Compare commits
1
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
  fredandscooby
|
b09589f02c |
@@ -34,7 +34,6 @@ export default tseslint.config(
|
||||
'packages/storage/vitest.config.ts',
|
||||
'packages/mosaic/vitest.config.ts',
|
||||
'packages/mosaic/__tests__/*.ts',
|
||||
'packages/forge/__tests__/*.ts',
|
||||
'tools/federation-harness/*.ts',
|
||||
],
|
||||
},
|
||||
|
||||
@@ -539,43 +539,3 @@ Not every brief needs full Board of Directors review. The classification system
|
||||
### Backward compatibility
|
||||
|
||||
Existing briefs without a `class` field are auto-classified. The default (no matching keywords) is `strategic`, so all existing runs get the full pipeline unless keywords trigger `technical`.
|
||||
|
||||
---
|
||||
|
||||
## Fail-Closed Execution & Explicit Simulation (SDLC-D-035)
|
||||
|
||||
**Added:** 2026-08-17
|
||||
|
||||
Forge fails closed when a required capability is missing. It never runs a
|
||||
pipeline with a stub executor and reports success.
|
||||
|
||||
### Normal mode (default)
|
||||
|
||||
- No task executor wired → the CLI exits nonzero with the typed capability
|
||||
error `FORGE_NO_EXECUTOR`. No run is created.
|
||||
- A stage whose gate is approval-based (board approval, planning approvals,
|
||||
remediation re-review, discovery/analysis attestations) records a typed
|
||||
`waiting-for-authority` stage result and raises `FORGE_AUTHORITY_REQUIRED`.
|
||||
It never passes vacuously.
|
||||
- A stage whose gate requires an unwired provider (AI reviewer, CI pipeline)
|
||||
records a typed `blocked` stage result and raises `FORGE_NO_REVIEWER` /
|
||||
`FORGE_NO_CI_PIPELINE`. The synthetic echo-review approval in `06-review`
|
||||
and all vacuous `true` gates were removed.
|
||||
|
||||
### Explicit simulation (`--simulate`)
|
||||
|
||||
Opts into stub/synthetic execution. Every stage result, every gate result, and
|
||||
the run manifest carry the distinct typed status `simulated` (manifest also
|
||||
records `mode: "simulated"`). `simulated` is a non-satisfying outcome:
|
||||
`isSatisfyingOutcome()` and all completion/gate consumers treat only `passed`
|
||||
as satisfying. The CLI exits 0 for a simulated run only because the caller
|
||||
explicitly passed `--simulate`, and prints a loud SIMULATED banner.
|
||||
|
||||
### Typed outcome model
|
||||
|
||||
Every gate/task outcome is one of the closed set
|
||||
`passed | failed | blocked | error | waiting-for-authority | simulated |
|
||||
not-applicable`, with the reason recorded on the stage status and each gate
|
||||
result in `manifest.json`. Missing implementations, missing gate evidence,
|
||||
unknown stages, process errors, and timeouts map to fail-closed members —
|
||||
never to `passed`.
|
||||
|
||||
@@ -1,319 +0,0 @@
|
||||
import fs from 'node:fs';
|
||||
import os from 'node:os';
|
||||
import path from 'node:path';
|
||||
import { describe, it, expect, beforeEach, afterEach } from 'vitest';
|
||||
|
||||
import { generateBoardTasks } from '../src/board-tasks.js';
|
||||
import { STAGE_SPECS } from '../src/constants.js';
|
||||
import { ForgeCapabilityError } from '../src/errors.js';
|
||||
import {
|
||||
evaluateStageGates,
|
||||
gateLabel,
|
||||
isCommandGate,
|
||||
isSatisfyingOutcome,
|
||||
} from '../src/outcomes.js';
|
||||
import { loadManifest, runPipeline } from '../src/pipeline-runner.js';
|
||||
import type { ForgeTask, ForgeTaskResult, TaskExecutor } from '../src/types.js';
|
||||
|
||||
/**
|
||||
* Mock real executor that returns typed results.
|
||||
*
|
||||
* Command gates are "verified" by the mock so normal-mode runs can pass
|
||||
* mechanically gated stages; authority/provider gates are never reported
|
||||
* because they have no mechanical implementation.
|
||||
*/
|
||||
function createTypedExecutor(options?: {
|
||||
failStage?: string;
|
||||
gateOutcomes?: Record<string, 'passed' | 'failed' | 'simulated' | 'error' | 'blocked'>;
|
||||
}): TaskExecutor & { submittedTasks: ForgeTask[] } {
|
||||
const submittedTasks: ForgeTask[] = [];
|
||||
return {
|
||||
submittedTasks,
|
||||
async submitTask(task: ForgeTask) {
|
||||
submittedTasks.push(task);
|
||||
},
|
||||
async waitForCompletion(taskId: string): Promise<ForgeTaskResult> {
|
||||
const task = submittedTasks.find((t) => t.id === taskId);
|
||||
const stageName = task?.metadata?.['stageName'] as string | undefined;
|
||||
|
||||
if (options?.failStage && stageName === options.failStage) {
|
||||
return {
|
||||
task_id: taskId,
|
||||
outcome: 'failed',
|
||||
reason: 'mock task failure',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 1,
|
||||
gate_results: [],
|
||||
};
|
||||
}
|
||||
|
||||
const gateResults = (task?.qualityGates ?? [])
|
||||
.filter((gate) => isCommandGate(gate))
|
||||
.map((gate) => {
|
||||
const label = gateLabel(gate);
|
||||
const outcome = options?.gateOutcomes?.[label] ?? 'passed';
|
||||
return {
|
||||
gate: label,
|
||||
outcome,
|
||||
reason: outcome === 'passed' ? 'mock verified' : `mock gate outcome: ${outcome}`,
|
||||
};
|
||||
});
|
||||
|
||||
return {
|
||||
task_id: taskId,
|
||||
outcome: 'passed',
|
||||
reason: 'mock verified',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 0,
|
||||
gate_results: gateResults,
|
||||
};
|
||||
},
|
||||
async getTaskStatus() {
|
||||
return 'completed' as const;
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
describe('fail-closed: no executor wired', () => {
|
||||
let tmpDir: string;
|
||||
let briefPath: string;
|
||||
|
||||
beforeEach(() => {
|
||||
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-failclosed-'));
|
||||
briefPath = path.join(tmpDir, 'brief.md');
|
||||
fs.writeFileSync(briefPath, '# Fix bug\n\nA bugfix for lint cleanup.');
|
||||
});
|
||||
|
||||
afterEach(() => {
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true });
|
||||
});
|
||||
|
||||
it('throws a typed FORGE_NO_EXECUTOR capability error without --simulate', async () => {
|
||||
await expect(
|
||||
runPipeline(briefPath, tmpDir, {
|
||||
// no executor, no simulate — must fail closed, never run with a stub
|
||||
stages: ['00-intake'],
|
||||
}),
|
||||
).rejects.toMatchObject({
|
||||
name: 'ForgeCapabilityError',
|
||||
code: 'FORGE_NO_EXECUTOR',
|
||||
capability: 'task-executor',
|
||||
});
|
||||
});
|
||||
|
||||
it('does not create a run directory when failing closed on a missing executor', async () => {
|
||||
try {
|
||||
await runPipeline(briefPath, tmpDir, { stages: ['00-intake'] });
|
||||
} catch {
|
||||
// expected
|
||||
}
|
||||
expect(fs.existsSync(path.join(tmpDir, '.forge', 'runs'))).toBe(false);
|
||||
});
|
||||
|
||||
it('completes with every result typed simulated when simulate is set', async () => {
|
||||
const result = await runPipeline(briefPath, tmpDir, {
|
||||
simulate: true,
|
||||
stages: ['00-intake', '00b-discovery', '02-planning-1', '06-review'],
|
||||
});
|
||||
|
||||
expect(result.manifest.mode).toBe('simulated');
|
||||
expect(result.manifest.status).toBe('simulated');
|
||||
|
||||
for (const stage of result.stages) {
|
||||
const stageStatus = result.manifest.stages[stage];
|
||||
expect(stageStatus?.status, `stage ${stage}`).toBe('simulated');
|
||||
expect(stageStatus?.status, `stage ${stage}`).not.toBe('passed');
|
||||
expect(stageStatus?.reason, `stage ${stage}`).toBeTruthy();
|
||||
for (const gateResult of stageStatus?.gateResults ?? []) {
|
||||
expect(gateResult.outcome, `gate ${gateResult.gate} of ${stage}`).toBe('simulated');
|
||||
expect(gateResult.outcome, `gate ${gateResult.gate} of ${stage}`).not.toBe('passed');
|
||||
}
|
||||
}
|
||||
|
||||
// The persisted manifest agrees.
|
||||
const persisted = loadManifest(result.runDir);
|
||||
expect(persisted.mode).toBe('simulated');
|
||||
expect(persisted.status).toBe('simulated');
|
||||
expect(persisted.stages['02-planning-1']?.status).toBe('simulated');
|
||||
});
|
||||
});
|
||||
|
||||
describe('fail-closed: typed outcome model', () => {
|
||||
it('only passed satisfies the gate/dependency predicate', () => {
|
||||
expect(isSatisfyingOutcome('passed')).toBe(true);
|
||||
expect(isSatisfyingOutcome('failed')).toBe(false);
|
||||
expect(isSatisfyingOutcome('blocked')).toBe(false);
|
||||
expect(isSatisfyingOutcome('error')).toBe(false);
|
||||
expect(isSatisfyingOutcome('waiting-for-authority')).toBe(false);
|
||||
expect(isSatisfyingOutcome('simulated')).toBe(false);
|
||||
expect(isSatisfyingOutcome('not-applicable')).toBe(false);
|
||||
});
|
||||
|
||||
it('a simulated gate result cannot satisfy the stage gate evaluation', () => {
|
||||
const evaluation = evaluateStageGates('05-coding', STAGE_SPECS['05-coding']!.qualityGates, {
|
||||
task_id: 'FORGE-x-05',
|
||||
outcome: 'passed',
|
||||
reason: 'executor claims success',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 0,
|
||||
gate_results: [{ gate: 'pnpm lint', outcome: 'simulated', reason: 'simulated gate' }],
|
||||
});
|
||||
expect(isSatisfyingOutcome(evaluation.outcome)).toBe(false);
|
||||
expect(evaluation.outcome).toBe('error');
|
||||
});
|
||||
|
||||
it('a simulated task outcome cannot satisfy evaluation in normal mode', () => {
|
||||
const evaluation = evaluateStageGates('00-intake', [], {
|
||||
task_id: 'FORGE-x-00',
|
||||
outcome: 'simulated',
|
||||
reason: 'executor reported simulated',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 0,
|
||||
gate_results: [],
|
||||
});
|
||||
expect(isSatisfyingOutcome(evaluation.outcome)).toBe(false);
|
||||
});
|
||||
|
||||
it('a missing gate result blocks the stage instead of passing vacuously', () => {
|
||||
const evaluation = evaluateStageGates('05-coding', STAGE_SPECS['05-coding']!.qualityGates, {
|
||||
task_id: 'FORGE-x-05',
|
||||
outcome: 'passed',
|
||||
reason: 'executor claims success',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 0,
|
||||
gate_results: [],
|
||||
});
|
||||
expect(evaluation.outcome).toBe('blocked');
|
||||
});
|
||||
});
|
||||
|
||||
describe('fail-closed: authority and provider gates', () => {
|
||||
let tmpDir: string;
|
||||
let briefPath: string;
|
||||
|
||||
beforeEach(() => {
|
||||
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-authority-'));
|
||||
briefPath = path.join(tmpDir, 'brief.md');
|
||||
fs.writeFileSync(briefPath, '# Fix bug\n\nA bugfix for lint cleanup.');
|
||||
});
|
||||
|
||||
afterEach(() => {
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true });
|
||||
});
|
||||
|
||||
it.each(['02-planning-1', '03-planning-2', '04-planning-3', '07-remediate'])(
|
||||
'planning/remediation stage %s yields waiting-for-authority (not passed) in normal mode',
|
||||
async (stage) => {
|
||||
const executor = createTypedExecutor();
|
||||
let runDir: string | undefined;
|
||||
|
||||
try {
|
||||
await runPipeline(briefPath, tmpDir, {
|
||||
executor,
|
||||
stages: [stage as string],
|
||||
});
|
||||
expect.unreachable('runPipeline should have failed closed');
|
||||
} catch (err) {
|
||||
expect(err).toBeInstanceOf(ForgeCapabilityError);
|
||||
expect((err as ForgeCapabilityError).code).toBe('FORGE_AUTHORITY_REQUIRED');
|
||||
runDir = path.join(tmpDir, '.forge', 'runs');
|
||||
}
|
||||
|
||||
const runIds = fs.readdirSync(runDir!);
|
||||
expect(runIds).toHaveLength(1);
|
||||
const manifest = loadManifest(path.join(runDir!, runIds[0]!));
|
||||
expect(manifest.stages[stage]?.status).toBe('waiting-for-authority');
|
||||
expect(manifest.stages[stage]?.status).not.toBe('passed');
|
||||
expect(manifest.status).toBe('waiting-for-authority');
|
||||
},
|
||||
);
|
||||
|
||||
it('review stage fails closed with a typed FORGE_NO_REVIEWER error in normal mode', async () => {
|
||||
const executor = createTypedExecutor();
|
||||
|
||||
try {
|
||||
await runPipeline(briefPath, tmpDir, {
|
||||
executor,
|
||||
stages: ['06-review'],
|
||||
});
|
||||
expect.unreachable('runPipeline should have failed closed');
|
||||
} catch (err) {
|
||||
expect(err).toBeInstanceOf(ForgeCapabilityError);
|
||||
expect((err as ForgeCapabilityError).code).toBe('FORGE_NO_REVIEWER');
|
||||
expect((err as ForgeCapabilityError).capability).toBe('reviewer');
|
||||
}
|
||||
|
||||
const runsDir = path.join(tmpDir, '.forge', 'runs');
|
||||
const runIds = fs.readdirSync(runsDir);
|
||||
const manifest = loadManifest(path.join(runsDir, runIds[0]!));
|
||||
expect(manifest.stages['06-review']?.status).toBe('blocked');
|
||||
expect(manifest.stages['06-review']?.status).not.toBe('passed');
|
||||
expect(manifest.status).toBe('failed');
|
||||
});
|
||||
|
||||
it('review stage produces simulated results under --simulate', async () => {
|
||||
const result = await runPipeline(briefPath, tmpDir, {
|
||||
simulate: true,
|
||||
stages: ['06-review'],
|
||||
});
|
||||
|
||||
expect(result.manifest.mode).toBe('simulated');
|
||||
expect(result.manifest.stages['06-review']?.status).toBe('simulated');
|
||||
for (const gateResult of result.manifest.stages['06-review']?.gateResults ?? []) {
|
||||
expect(gateResult.outcome).toBe('simulated');
|
||||
}
|
||||
});
|
||||
|
||||
it('deploy stage fails closed without a wired ci-pipeline provider in normal mode', async () => {
|
||||
const executor = createTypedExecutor();
|
||||
|
||||
await expect(
|
||||
runPipeline(briefPath, tmpDir, {
|
||||
executor,
|
||||
stages: ['09-deploy'],
|
||||
}),
|
||||
).rejects.toMatchObject({
|
||||
name: 'ForgeCapabilityError',
|
||||
code: 'FORGE_NO_CI_PIPELINE',
|
||||
});
|
||||
});
|
||||
});
|
||||
|
||||
describe('fail-closed: no vacuous gate commands remain', () => {
|
||||
it('stage constants contain no echo/synthetic-approval, vacuous true, or empty gate commands', () => {
|
||||
for (const [stageName, spec] of Object.entries(STAGE_SPECS)) {
|
||||
for (const gate of spec.qualityGates) {
|
||||
const serialized = JSON.stringify(gate);
|
||||
// The echo-review synthetic approval must be gone.
|
||||
expect(serialized, `stage ${stageName} gate ${serialized}`).not.toContain('echo');
|
||||
expect(serialized, `stage ${stageName} gate ${serialized}`).not.toMatch(/"verdict"\s*:/);
|
||||
expect(serialized, `stage ${stageName} gate ${serialized}`).not.toMatch(
|
||||
/"summary"\s*:\s*"review-pass"/,
|
||||
);
|
||||
// No vacuous literal `true` gate.
|
||||
expect(gate, `stage ${stageName}`).not.toBe('true');
|
||||
// Command gates must carry a real, non-empty command.
|
||||
if (isCommandGate(gate)) {
|
||||
const command = typeof gate === 'string' ? gate : gate.command;
|
||||
expect(command.trim().length, `stage ${stageName} gate ${serialized}`).toBeGreaterThan(0);
|
||||
}
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
it('board tasks contain no vacuous true gates', () => {
|
||||
const tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-board-gates-'));
|
||||
try {
|
||||
const tasks = generateBoardTasks('# Brief', [], tmpDir, 'BOARD-TEST');
|
||||
for (const task of tasks) {
|
||||
for (const gate of task.qualityGates) {
|
||||
expect(gate, `task ${task.id}`).not.toBe('true');
|
||||
const serialized = JSON.stringify(gate);
|
||||
expect(serialized, `task ${task.id} gate ${serialized}`).not.toContain('echo');
|
||||
}
|
||||
}
|
||||
} finally {
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true });
|
||||
}
|
||||
});
|
||||
});
|
||||
@@ -12,10 +12,10 @@ import {
|
||||
resumePipeline,
|
||||
getPipelineStatus,
|
||||
} from '../src/pipeline-runner.js';
|
||||
import type { ForgeTask, ForgeTaskResult, RunManifest, TaskExecutor } from '../src/types.js';
|
||||
import { gateLabel, isCommandGate } from '../src/outcomes.js';
|
||||
import type { ForgeTask, RunManifest, TaskExecutor } from '../src/types.js';
|
||||
import type { TaskResult } from '@mosaicstack/macp';
|
||||
|
||||
/** Mock TaskExecutor that records submitted tasks and returns typed results. */
|
||||
/** Mock TaskExecutor that records submitted tasks and returns success. */
|
||||
function createMockExecutor(options?: {
|
||||
failStage?: string;
|
||||
}): TaskExecutor & { submittedTasks: ForgeTask[] } {
|
||||
@@ -25,7 +25,7 @@ function createMockExecutor(options?: {
|
||||
async submitTask(task: ForgeTask) {
|
||||
submittedTasks.push(task);
|
||||
},
|
||||
async waitForCompletion(taskId: string): Promise<ForgeTaskResult> {
|
||||
async waitForCompletion(taskId: string): Promise<TaskResult> {
|
||||
const failStage = options?.failStage;
|
||||
const task = submittedTasks.find((t) => t.id === taskId);
|
||||
const stageName = task?.metadata?.['stageName'] as string | undefined;
|
||||
@@ -33,8 +33,7 @@ function createMockExecutor(options?: {
|
||||
if (failStage && stageName === failStage) {
|
||||
return {
|
||||
task_id: taskId,
|
||||
outcome: 'failed',
|
||||
reason: 'mock task failure',
|
||||
status: 'failed',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 1,
|
||||
gate_results: [],
|
||||
@@ -42,17 +41,10 @@ function createMockExecutor(options?: {
|
||||
}
|
||||
return {
|
||||
task_id: taskId,
|
||||
outcome: 'passed',
|
||||
reason: 'mock verified',
|
||||
status: 'completed',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 0,
|
||||
gate_results: (task?.qualityGates ?? [])
|
||||
.filter((gate) => isCommandGate(gate))
|
||||
.map((gate) => ({
|
||||
gate: gateLabel(gate),
|
||||
outcome: 'passed' as const,
|
||||
reason: 'mock verified',
|
||||
})),
|
||||
gate_results: [],
|
||||
};
|
||||
},
|
||||
async getTaskStatus() {
|
||||
@@ -164,13 +156,12 @@ describe('runPipeline', () => {
|
||||
const executor = createMockExecutor();
|
||||
const result = await runPipeline(briefPath, tmpDir, {
|
||||
executor,
|
||||
stages: ['00-intake', '05-coding'],
|
||||
stages: ['00-intake', '00b-discovery'],
|
||||
});
|
||||
|
||||
expect(result.runId).toMatch(/^\d{8}-\d{6}$/);
|
||||
expect(result.stages).toEqual(['00-intake', '05-coding']);
|
||||
expect(result.stages).toEqual(['00-intake', '00b-discovery']);
|
||||
expect(result.manifest.status).toBe('completed');
|
||||
expect(result.manifest.mode).toBe('normal');
|
||||
expect(executor.submittedTasks).toHaveLength(2);
|
||||
});
|
||||
|
||||
@@ -189,17 +180,12 @@ describe('runPipeline', () => {
|
||||
const executor = createMockExecutor();
|
||||
const result = await runPipeline(briefPath, tmpDir, {
|
||||
executor,
|
||||
stages: ['00-intake', '05-coding'],
|
||||
stages: ['00-intake', '00b-discovery'],
|
||||
});
|
||||
|
||||
const manifest = loadManifest(result.runDir);
|
||||
expect(manifest.stages['00-intake']?.status).toBe('passed');
|
||||
expect(manifest.stages['05-coding']?.status).toBe('passed');
|
||||
expect(manifest.stages['05-coding']?.gateResults?.map((g) => g.outcome)).toEqual([
|
||||
'passed',
|
||||
'passed',
|
||||
'passed',
|
||||
]);
|
||||
expect(manifest.stages['00b-discovery']?.status).toBe('passed');
|
||||
});
|
||||
|
||||
it('respects CLI class override', async () => {
|
||||
@@ -229,7 +215,7 @@ describe('runPipeline', () => {
|
||||
const executor = createMockExecutor();
|
||||
await runPipeline(briefPath, tmpDir, {
|
||||
executor,
|
||||
stages: ['00-intake', '05-coding', '08-test'],
|
||||
stages: ['00-intake', '00b-discovery', '02-planning-1'],
|
||||
});
|
||||
|
||||
expect(executor.submittedTasks[0]!.dependsOn).toBeUndefined();
|
||||
@@ -238,14 +224,14 @@ describe('runPipeline', () => {
|
||||
});
|
||||
|
||||
it('handles stage failure', async () => {
|
||||
const executor = createMockExecutor({ failStage: '05-coding' });
|
||||
const executor = createMockExecutor({ failStage: '00b-discovery' });
|
||||
|
||||
await expect(
|
||||
runPipeline(briefPath, tmpDir, {
|
||||
executor,
|
||||
stages: ['00-intake', '05-coding'],
|
||||
stages: ['00-intake', '00b-discovery'],
|
||||
}),
|
||||
).rejects.toThrow('Stage 05-coding failed');
|
||||
).rejects.toThrow('Stage 00b-discovery failed');
|
||||
});
|
||||
|
||||
it('marks manifest as failed on stage failure', async () => {
|
||||
@@ -284,143 +270,30 @@ describe('resumePipeline', () => {
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true });
|
||||
});
|
||||
|
||||
it('resumes from first incomplete stage and fails closed at the next provider gate', async () => {
|
||||
// Simulate a run whose authority stages were approved out-of-band
|
||||
// (recorded as passed) and whose coding stage failed mechanically.
|
||||
const runId = '20260101-000000';
|
||||
const runDir = path.join(tmpDir, '.forge', 'runs', runId);
|
||||
fs.mkdirSync(runDir, { recursive: true });
|
||||
const passed = { status: 'passed' as const, startedAt: '2026-01-01T00:00:00Z' };
|
||||
saveManifest(runDir, {
|
||||
runId,
|
||||
brief: briefPath,
|
||||
codebase: tmpDir,
|
||||
briefClass: 'hotfix',
|
||||
classSource: 'frontmatter',
|
||||
forceBoard: false,
|
||||
mode: 'normal',
|
||||
createdAt: '2026-01-01T00:00:00Z',
|
||||
updatedAt: '2026-01-01T00:00:00Z',
|
||||
currentStage: '05-coding',
|
||||
status: 'failed',
|
||||
stages: {
|
||||
'00-intake': passed,
|
||||
'00b-discovery': passed,
|
||||
'02-planning-1': passed,
|
||||
'03-planning-2': passed,
|
||||
'04-planning-3': passed,
|
||||
'05-coding': { status: 'failed', reason: 'gate failed' },
|
||||
},
|
||||
it('resumes from first incomplete stage', async () => {
|
||||
// First run fails on discovery
|
||||
const executor1 = createMockExecutor({ failStage: '00b-discovery' });
|
||||
let runDir: string;
|
||||
|
||||
try {
|
||||
await runPipeline(briefPath, tmpDir, {
|
||||
executor: executor1,
|
||||
stages: ['00-intake', '00b-discovery', '02-planning-1'],
|
||||
});
|
||||
|
||||
// Resume re-runs 05-coding (the first non-passed stage), then fails
|
||||
// closed at 06-review because no reviewer provider is wired.
|
||||
const executor = createMockExecutor();
|
||||
await expect(resumePipeline(runDir, executor)).rejects.toMatchObject({
|
||||
name: 'ForgeCapabilityError',
|
||||
code: 'FORGE_NO_REVIEWER',
|
||||
});
|
||||
|
||||
const manifest = loadManifest(runDir);
|
||||
expect(manifest.stages['05-coding']?.status).toBe('passed');
|
||||
expect(manifest.stages['06-review']?.status).toBe('blocked');
|
||||
expect(manifest.status).toBe('failed');
|
||||
});
|
||||
|
||||
it('resumes to completion as simulated under explicit simulate', async () => {
|
||||
const runId = '20260101-000003';
|
||||
const runDir = path.join(tmpDir, '.forge', 'runs', runId);
|
||||
fs.mkdirSync(runDir, { recursive: true });
|
||||
const passed = { status: 'passed' as const, startedAt: '2026-01-01T00:00:00Z' };
|
||||
saveManifest(runDir, {
|
||||
runId,
|
||||
brief: briefPath,
|
||||
codebase: tmpDir,
|
||||
briefClass: 'hotfix',
|
||||
classSource: 'frontmatter',
|
||||
forceBoard: false,
|
||||
mode: 'normal',
|
||||
createdAt: '2026-01-01T00:00:00Z',
|
||||
updatedAt: '2026-01-01T00:00:00Z',
|
||||
currentStage: '05-coding',
|
||||
status: 'failed',
|
||||
stages: {
|
||||
'00-intake': passed,
|
||||
'00b-discovery': passed,
|
||||
'02-planning-1': passed,
|
||||
'03-planning-2': passed,
|
||||
'04-planning-3': passed,
|
||||
'05-coding': { status: 'failed', reason: 'gate failed' },
|
||||
},
|
||||
});
|
||||
|
||||
const result = await resumePipeline(runDir, undefined, { simulate: true });
|
||||
|
||||
expect(result.manifest.status).toBe('simulated');
|
||||
expect(result.manifest.mode).toBe('simulated');
|
||||
expect(result.stages[0]).toBe('05-coding');
|
||||
for (const stage of result.stages) {
|
||||
expect(result.manifest.stages[stage]?.status).toBe('simulated');
|
||||
} catch {
|
||||
// expected
|
||||
}
|
||||
});
|
||||
|
||||
it('fails closed on resume when the next stage needs authority sign-off', async () => {
|
||||
const runId = '20260101-000001';
|
||||
const runDir = path.join(tmpDir, '.forge', 'runs', runId);
|
||||
fs.mkdirSync(runDir, { recursive: true });
|
||||
saveManifest(runDir, {
|
||||
runId,
|
||||
brief: briefPath,
|
||||
codebase: tmpDir,
|
||||
briefClass: 'hotfix',
|
||||
classSource: 'frontmatter',
|
||||
forceBoard: false,
|
||||
mode: 'normal',
|
||||
createdAt: '2026-01-01T00:00:00Z',
|
||||
updatedAt: '2026-01-01T00:00:00Z',
|
||||
currentStage: '00-intake',
|
||||
status: 'in_progress',
|
||||
stages: {
|
||||
'00-intake': { status: 'passed' },
|
||||
},
|
||||
});
|
||||
const runsDir = path.join(tmpDir, '.forge', 'runs');
|
||||
runDir = path.join(runsDir, fs.readdirSync(runsDir)[0]!);
|
||||
|
||||
const executor = createMockExecutor();
|
||||
await expect(resumePipeline(runDir, executor)).rejects.toMatchObject({
|
||||
name: 'ForgeCapabilityError',
|
||||
code: 'FORGE_AUTHORITY_REQUIRED',
|
||||
});
|
||||
// Resume should pick up from 00b-discovery
|
||||
const executor2 = createMockExecutor();
|
||||
const result = await resumePipeline(runDir, executor2);
|
||||
|
||||
const manifest = loadManifest(runDir);
|
||||
expect(manifest.stages['00b-discovery']?.status).toBe('waiting-for-authority');
|
||||
expect(manifest.status).toBe('waiting-for-authority');
|
||||
});
|
||||
|
||||
it('fails closed on resume without an executor or --simulate', async () => {
|
||||
const runId = '20260101-000002';
|
||||
const runDir = path.join(tmpDir, '.forge', 'runs', runId);
|
||||
fs.mkdirSync(runDir, { recursive: true });
|
||||
saveManifest(runDir, {
|
||||
runId,
|
||||
brief: briefPath,
|
||||
codebase: tmpDir,
|
||||
briefClass: 'hotfix',
|
||||
classSource: 'frontmatter',
|
||||
forceBoard: false,
|
||||
mode: 'normal',
|
||||
createdAt: '2026-01-01T00:00:00Z',
|
||||
updatedAt: '2026-01-01T00:00:00Z',
|
||||
currentStage: '00-intake',
|
||||
status: 'in_progress',
|
||||
stages: {
|
||||
'00-intake': { status: 'passed' },
|
||||
},
|
||||
});
|
||||
|
||||
await expect(resumePipeline(runDir)).rejects.toMatchObject({
|
||||
name: 'ForgeCapabilityError',
|
||||
code: 'FORGE_NO_EXECUTOR',
|
||||
});
|
||||
expect(result.manifest.status).toBe('completed');
|
||||
// Should have re-run from 00b-discovery onward
|
||||
expect(result.stages[0]).toBe('00b-discovery');
|
||||
});
|
||||
});
|
||||
|
||||
|
||||
@@ -95,14 +95,7 @@ export function generateBoardTasks(
|
||||
briefPath,
|
||||
resultPath: resultRelPath,
|
||||
timeoutSeconds: 120,
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'board-approval',
|
||||
reason:
|
||||
'persona evaluation is judged by board synthesis (authority review); no mechanical gate exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
metadata: {
|
||||
personaName: persona.name,
|
||||
personaSlug: persona.slug,
|
||||
@@ -128,13 +121,7 @@ export function generateBoardTasks(
|
||||
timeoutSeconds: 120,
|
||||
dependsOn: personaTaskIds,
|
||||
dependsOnPolicy: 'all_terminal',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'board-approval',
|
||||
reason: 'board synthesis is an authority decision; no mechanical gate exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
metadata: {
|
||||
resultOutputPath: synthesisResult,
|
||||
inputResultPaths: personaResultPaths,
|
||||
|
||||
@@ -1,11 +1,7 @@
|
||||
import fs from 'node:fs';
|
||||
import os from 'node:os';
|
||||
import path from 'node:path';
|
||||
import { Command } from 'commander';
|
||||
import { describe, expect, it, vi, beforeEach, afterEach } from 'vitest';
|
||||
import { describe, expect, it } from 'vitest';
|
||||
|
||||
import { registerForgeCommand } from './cli.js';
|
||||
import { loadManifest } from './pipeline-runner.js';
|
||||
|
||||
describe('registerForgeCommand', () => {
|
||||
it('registers a "forge" command on the parent program', () => {
|
||||
@@ -59,94 +55,3 @@ describe('registerForgeCommand', () => {
|
||||
}).not.toThrow();
|
||||
});
|
||||
});
|
||||
|
||||
describe('forge run fail-closed behavior (SDLC-D-035)', () => {
|
||||
let tmpDir: string;
|
||||
let briefPath: string;
|
||||
let errSpy: ReturnType<typeof vi.spyOn>;
|
||||
let logSpy: ReturnType<typeof vi.spyOn>;
|
||||
let prevExitCode: string | number | null | undefined;
|
||||
|
||||
const parse = (args: string[]) => {
|
||||
const program = new Command();
|
||||
registerForgeCommand(program);
|
||||
return program.parseAsync(['forge', ...args], { from: 'user' });
|
||||
};
|
||||
|
||||
beforeEach(() => {
|
||||
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'forge-cli-failclosed-'));
|
||||
briefPath = path.join(tmpDir, 'brief.md');
|
||||
fs.writeFileSync(briefPath, '# Fix bug\n\nA bugfix for lint cleanup.');
|
||||
errSpy = vi.spyOn(console, 'error').mockImplementation(() => {});
|
||||
logSpy = vi.spyOn(console, 'log').mockImplementation(() => {});
|
||||
prevExitCode = process.exitCode;
|
||||
});
|
||||
|
||||
afterEach(() => {
|
||||
errSpy.mockRestore();
|
||||
logSpy.mockRestore();
|
||||
process.exitCode = prevExitCode;
|
||||
fs.rmSync(tmpDir, { recursive: true, force: true });
|
||||
});
|
||||
|
||||
it('exits nonzero with a typed FORGE_NO_EXECUTOR error when no executor is wired and --simulate is absent', async () => {
|
||||
await parse(['run', '--brief', briefPath, '--codebase', tmpDir]);
|
||||
|
||||
expect(process.exitCode).toBe(1);
|
||||
const errText = errSpy.mock.calls.map((c) => c.join(' ')).join('\n');
|
||||
expect(errText).toContain('FORGE_NO_EXECUTOR');
|
||||
// It must never run the pipeline with a stub and report success.
|
||||
expect(fs.existsSync(path.join(tmpDir, '.forge', 'runs'))).toBe(false);
|
||||
});
|
||||
|
||||
it('completes with typed simulated results and exit 0 under explicit --simulate', async () => {
|
||||
await parse(['run', '--brief', briefPath, '--codebase', tmpDir, '--simulate']);
|
||||
|
||||
expect(process.exitCode).toBeUndefined();
|
||||
|
||||
// Loud simulated-mode summary.
|
||||
const logText = logSpy.mock.calls.map((c) => c.join(' ')).join('\n');
|
||||
expect(logText).toContain('SIMULATED');
|
||||
|
||||
// Manifest records the mode and simulated per-result statuses.
|
||||
const runsDir = path.join(tmpDir, '.forge', 'runs');
|
||||
const runIds = fs.readdirSync(runsDir);
|
||||
expect(runIds).toHaveLength(1);
|
||||
const manifest = loadManifest(path.join(runsDir, runIds[0]!));
|
||||
expect(manifest.mode).toBe('simulated');
|
||||
expect(manifest.status).toBe('simulated');
|
||||
for (const stageStatus of Object.values(manifest.stages)) {
|
||||
expect(stageStatus?.status).toBe('simulated');
|
||||
for (const gateResult of stageStatus?.gateResults ?? []) {
|
||||
expect(gateResult.outcome).toBe('simulated');
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
it('resume exits nonzero with a typed FORGE_NO_EXECUTOR error without --simulate', async () => {
|
||||
const runDir = path.join(tmpDir, '.forge', 'runs', '20260101-000000');
|
||||
fs.mkdirSync(runDir, { recursive: true });
|
||||
fs.writeFileSync(
|
||||
path.join(runDir, 'manifest.json'),
|
||||
JSON.stringify({
|
||||
runId: '20260101-000000',
|
||||
brief: briefPath,
|
||||
codebase: tmpDir,
|
||||
briefClass: 'hotfix',
|
||||
classSource: 'frontmatter',
|
||||
forceBoard: false,
|
||||
createdAt: '2026-01-01T00:00:00Z',
|
||||
updatedAt: '2026-01-01T00:00:00Z',
|
||||
currentStage: '00-intake',
|
||||
status: 'in_progress',
|
||||
stages: { '00-intake': { status: 'passed' } },
|
||||
}),
|
||||
);
|
||||
|
||||
await parse(['resume', '20260101-000000', '--project', tmpDir]);
|
||||
|
||||
expect(process.exitCode).toBe(1);
|
||||
const errText = errSpy.mock.calls.map((c) => c.join(' ')).join('\n');
|
||||
expect(errText).toContain('FORGE_NO_EXECUTOR');
|
||||
});
|
||||
});
|
||||
|
||||
+47
-121
@@ -5,47 +5,37 @@ import type { Command } from 'commander';
|
||||
|
||||
import { classifyBrief } from './brief-classifier.js';
|
||||
import { STAGE_LABELS, STAGE_SEQUENCE } from './constants.js';
|
||||
import { ForgeCapabilityError } from './errors.js';
|
||||
import { getEffectivePersonas, loadBoardPersonas } from './persona-loader.js';
|
||||
import { generateRunId, getPipelineStatus, loadManifest, runPipeline } from './pipeline-runner.js';
|
||||
import { createSimulatedExecutor } from './simulated-executor.js';
|
||||
import type { PipelineOptions, RunManifest, RunMode } from './types.js';
|
||||
import type { PipelineOptions, RunManifest, TaskExecutor } from './types.js';
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Stub executor — used when no real executor is wired at CLI invocation time.
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
const stubExecutor: TaskExecutor = {
|
||||
async submitTask(task) {
|
||||
console.log(` [forge] stage submitted: ${task.id} (${task.title})`);
|
||||
},
|
||||
async waitForCompletion(taskId, _timeoutMs) {
|
||||
console.log(` [forge] stage complete: ${taskId}`);
|
||||
return {
|
||||
task_id: taskId,
|
||||
status: 'completed' as const,
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 0,
|
||||
gate_results: [],
|
||||
};
|
||||
},
|
||||
async getTaskStatus(_taskId) {
|
||||
return 'completed' as const;
|
||||
},
|
||||
};
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Helpers
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
/** Resolve a run's effective mode, defaulting legacy manifests to normal. */
|
||||
function runModeOf(manifest: RunManifest): RunMode {
|
||||
return manifest.mode ?? 'normal';
|
||||
}
|
||||
|
||||
/** Print a loud banner so a simulated run can never be misread as verified. */
|
||||
function printSimulatedBanner(): void {
|
||||
console.log('');
|
||||
console.log('[forge] ===============================================================');
|
||||
console.log('[forge] MODE: SIMULATED — no stage or gate was really executed.');
|
||||
console.log('[forge] All results are synthetic and MUST NOT be read as verified');
|
||||
console.log('[forge] success. Wire a real executor/providers and re-run to verify.');
|
||||
console.log('[forge] ===============================================================');
|
||||
}
|
||||
|
||||
/** Print a typed error line for fail-closed capability errors. */
|
||||
function printCapabilityError(err: ForgeCapabilityError): void {
|
||||
console.error(`[forge] error ${err.code}: ${err.message}`);
|
||||
console.error(`[forge] missing capability: ${err.capability}`);
|
||||
}
|
||||
|
||||
/** Handle a pipeline error uniformly: typed capability errors get their code. */
|
||||
function handlePipelineError(err: unknown): void {
|
||||
if (err instanceof ForgeCapabilityError) {
|
||||
printCapabilityError(err);
|
||||
} else {
|
||||
console.error(`[forge] pipeline failed: ${err instanceof Error ? err.message : String(err)}`);
|
||||
}
|
||||
process.exitCode = 1;
|
||||
}
|
||||
|
||||
function formatDuration(startedAt?: string, completedAt?: string): string {
|
||||
if (!startedAt || !completedAt) return '-';
|
||||
const ms = new Date(completedAt).getTime() - new Date(startedAt).getTime();
|
||||
@@ -54,24 +44,19 @@ function formatDuration(startedAt?: string, completedAt?: string): string {
|
||||
}
|
||||
|
||||
function printManifestTable(manifest: RunManifest): void {
|
||||
const mode = runModeOf(manifest);
|
||||
console.log(`\nRun ID : ${manifest.runId}`);
|
||||
console.log(`Status : ${manifest.status}`);
|
||||
console.log(`Mode : ${mode}`);
|
||||
if (mode === 'simulated') {
|
||||
console.log('WARNING: SIMULATED RUN — results are synthetic, not verified success.');
|
||||
}
|
||||
console.log(`Brief : ${manifest.brief}`);
|
||||
console.log(`Class : ${manifest.briefClass} (${manifest.classSource})`);
|
||||
console.log(`Updated: ${manifest.updatedAt}`);
|
||||
console.log('');
|
||||
console.log('Stage'.padEnd(22) + 'Status'.padEnd(24) + 'Duration');
|
||||
console.log('-'.repeat(60));
|
||||
console.log('Stage'.padEnd(22) + 'Status'.padEnd(14) + 'Duration');
|
||||
console.log('-'.repeat(50));
|
||||
for (const stage of STAGE_SEQUENCE) {
|
||||
const s = manifest.stages[stage];
|
||||
if (!s) continue;
|
||||
const label = (STAGE_LABELS[stage] ?? stage).padEnd(22);
|
||||
const status = s.status.padEnd(24);
|
||||
const status = s.status.padEnd(14);
|
||||
const dur = formatDuration(s.startedAt, s.completedAt);
|
||||
console.log(`${label}${status}${dur}`);
|
||||
}
|
||||
@@ -105,58 +90,23 @@ function listRecentRuns(projectRoot?: string): void {
|
||||
}
|
||||
|
||||
console.log('\nRecent runs:');
|
||||
console.log('Run ID'.padEnd(22) + 'Status'.padEnd(24) + 'Mode'.padEnd(12) + 'Brief');
|
||||
console.log('-'.repeat(80));
|
||||
console.log('Run ID'.padEnd(22) + 'Status'.padEnd(14) + 'Brief');
|
||||
console.log('-'.repeat(70));
|
||||
|
||||
for (const runId of entries) {
|
||||
const runDir = path.join(runsDir, runId);
|
||||
try {
|
||||
const manifest = loadManifest(runDir);
|
||||
const status = manifest.status.padEnd(24);
|
||||
const mode = runModeOf(manifest).padEnd(12);
|
||||
const status = manifest.status.padEnd(14);
|
||||
const brief = path.basename(manifest.brief);
|
||||
console.log(`${runId.padEnd(22)}${status}${mode}${brief}`);
|
||||
console.log(`${runId.padEnd(22)}${status}${brief}`);
|
||||
} catch {
|
||||
console.log(`${runId.padEnd(22)}${'(unreadable)'.padEnd(24)}`);
|
||||
console.log(`${runId.padEnd(22)}${'(unreadable)'.padEnd(14)}`);
|
||||
}
|
||||
}
|
||||
console.log('');
|
||||
}
|
||||
|
||||
/**
|
||||
* Apply the exit-code policy for a finished pipeline run (SDLC-D-035):
|
||||
*
|
||||
* - exit 0 only for a verified `completed` normal run, or for an overall
|
||||
* `simulated` run when the caller explicitly passed --simulate;
|
||||
* - anything else exits nonzero so it can never be read as success.
|
||||
*/
|
||||
function applyRunExitPolicy(result: { manifest: RunManifest; runDir: string }, simulate: boolean) {
|
||||
const { manifest } = result;
|
||||
|
||||
if (runModeOf(manifest) === 'simulated') {
|
||||
if (!simulate || manifest.status !== 'simulated') {
|
||||
console.error(
|
||||
'[forge] error FORGE_MODE_MISMATCH: run reports simulated results without an explicit, ' +
|
||||
'consistent --simulate request; refusing to report success.',
|
||||
);
|
||||
process.exitCode = 1;
|
||||
return;
|
||||
}
|
||||
printSimulatedBanner();
|
||||
console.log(`[forge] run directory: ${result.runDir}`);
|
||||
return; // exit 0 — the caller explicitly opted into simulation
|
||||
}
|
||||
|
||||
if (manifest.status !== 'completed') {
|
||||
console.error(`[forge] run did not complete: terminal status '${manifest.status}'`);
|
||||
process.exitCode = 1;
|
||||
return;
|
||||
}
|
||||
|
||||
console.log(`[forge] pipeline complete (mode: normal): ${manifest.runId}`);
|
||||
console.log(`[forge] run directory: ${result.runDir}`);
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Register function
|
||||
// ---------------------------------------------------------------------------
|
||||
@@ -179,11 +129,6 @@ export function registerForgeCommand(parent: Command): void {
|
||||
.option('--config <path>', 'Path to forge config file (.forge/config.yaml)')
|
||||
.option('--codebase <path>', 'Codebase root to pass to the pipeline', process.cwd())
|
||||
.option('--dry-run', 'Print planned stages without executing', false)
|
||||
.option(
|
||||
'--simulate',
|
||||
'Simulate execution without real providers (every result is typed simulated, never verified)',
|
||||
false,
|
||||
)
|
||||
.action(
|
||||
async (opts: {
|
||||
brief: string;
|
||||
@@ -192,7 +137,6 @@ export function registerForgeCommand(parent: Command): void {
|
||||
config?: string;
|
||||
codebase: string;
|
||||
dryRun: boolean;
|
||||
simulate: boolean;
|
||||
}) => {
|
||||
const briefPath = path.resolve(opts.brief);
|
||||
|
||||
@@ -205,22 +149,14 @@ export function registerForgeCommand(parent: Command): void {
|
||||
const briefContent = fs.readFileSync(briefPath, 'utf-8');
|
||||
const briefClass = classifyBrief(briefContent);
|
||||
const projectRoot = opts.codebase;
|
||||
// A real executor is never wired at CLI invocation time today, so the
|
||||
// only executor we may construct is the explicitly-requested simulated
|
||||
// one. Normal mode fails closed with FORGE_NO_EXECUTOR.
|
||||
const executor = opts.simulate ? createSimulatedExecutor() : undefined;
|
||||
|
||||
if (opts.resume) {
|
||||
const runId = opts.runId ?? generateRunId();
|
||||
const runDir = resolveRunDir(runId, projectRoot);
|
||||
console.log(`[forge] resuming run: ${runId}`);
|
||||
try {
|
||||
const { resumePipeline } = await import('./pipeline-runner.js');
|
||||
const result = await resumePipeline(runDir, executor, { simulate: opts.simulate });
|
||||
applyRunExitPolicy(result, opts.simulate);
|
||||
} catch (err) {
|
||||
handlePipelineError(err);
|
||||
}
|
||||
const result = await resumePipeline(runDir, stubExecutor);
|
||||
console.log(`[forge] pipeline complete: ${result.runId}`);
|
||||
return;
|
||||
}
|
||||
|
||||
@@ -228,8 +164,7 @@ export function registerForgeCommand(parent: Command): void {
|
||||
briefClass,
|
||||
codebase: projectRoot,
|
||||
dryRun: opts.dryRun,
|
||||
executor,
|
||||
simulate: opts.simulate,
|
||||
executor: stubExecutor,
|
||||
};
|
||||
|
||||
if (opts.dryRun) {
|
||||
@@ -245,15 +180,16 @@ export function registerForgeCommand(parent: Command): void {
|
||||
|
||||
console.log(`[forge] starting pipeline for brief: ${briefPath}`);
|
||||
console.log(`[forge] classified as: ${briefClass}`);
|
||||
if (opts.simulate) {
|
||||
console.log('[forge] mode: SIMULATED (explicit --simulate)');
|
||||
}
|
||||
|
||||
try {
|
||||
const result = await runPipeline(briefPath, projectRoot, pipelineOptions);
|
||||
applyRunExitPolicy(result, opts.simulate);
|
||||
console.log(`[forge] pipeline complete: ${result.runId}`);
|
||||
console.log(`[forge] run directory: ${result.runDir}`);
|
||||
} catch (err) {
|
||||
handlePipelineError(err);
|
||||
console.error(
|
||||
`[forge] pipeline failed: ${err instanceof Error ? err.message : String(err)}`,
|
||||
);
|
||||
process.exitCode = 1;
|
||||
}
|
||||
},
|
||||
);
|
||||
@@ -288,12 +224,7 @@ export function registerForgeCommand(parent: Command): void {
|
||||
.command('resume <runId>')
|
||||
.description('Resume a stopped or failed pipeline run')
|
||||
.option('--project <path>', 'Project root (defaults to cwd)', process.cwd())
|
||||
.option(
|
||||
'--simulate',
|
||||
'Simulate execution without real providers (every result is typed simulated, never verified)',
|
||||
false,
|
||||
)
|
||||
.action(async (runId: string, opts: { project: string; simulate: boolean }) => {
|
||||
.action(async (runId: string, opts: { project: string }) => {
|
||||
const runDir = resolveRunDir(runId, opts.project);
|
||||
|
||||
if (!fs.existsSync(runDir)) {
|
||||
@@ -303,20 +234,15 @@ export function registerForgeCommand(parent: Command): void {
|
||||
}
|
||||
|
||||
console.log(`[forge] resuming run: ${runId}`);
|
||||
if (opts.simulate) {
|
||||
console.log('[forge] mode: SIMULATED (explicit --simulate)');
|
||||
}
|
||||
|
||||
// No real executor is wired at CLI invocation time; only the explicitly
|
||||
// requested simulated executor may be constructed (fail closed otherwise).
|
||||
const executor = opts.simulate ? createSimulatedExecutor() : undefined;
|
||||
|
||||
try {
|
||||
const { resumePipeline } = await import('./pipeline-runner.js');
|
||||
const result = await resumePipeline(runDir, executor, { simulate: opts.simulate });
|
||||
applyRunExitPolicy(result, opts.simulate);
|
||||
const result = await resumePipeline(runDir, stubExecutor);
|
||||
console.log(`[forge] pipeline complete: ${result.runId}`);
|
||||
console.log(`[forge] run directory: ${result.runDir}`);
|
||||
} catch (err) {
|
||||
handlePipelineError(err);
|
||||
console.error(`[forge] resume failed: ${err instanceof Error ? err.message : String(err)}`);
|
||||
process.exitCode = 1;
|
||||
}
|
||||
});
|
||||
|
||||
|
||||
@@ -9,16 +9,7 @@ export const PACKAGE_ROOT = path.resolve(path.dirname(fileURLToPath(import.meta.
|
||||
/** Pipeline asset directory (stages, agents, rails, gates, templates). */
|
||||
export const PIPELINE_DIR = path.join(PACKAGE_ROOT, 'pipeline');
|
||||
|
||||
/** Stage specifications — defines every pipeline stage.
|
||||
*\n * Gate semantics (SDLC-D-035): every gate is one of
|
||||
* - a real command string / GateEntry a mechanical runner can execute,
|
||||
* - an `authority` gate (human/board sign-off; produces waiting-for-authority),
|
||||
* - a `provider` gate (requires a wired provider such as a reviewer or CI pipeline).
|
||||
*
|
||||
* Vacuous gates (`true`, echo'd synthetic approvals, placeholder ci-pipeline
|
||||
* commands) are forbidden: a stage whose gate has no real implementation
|
||||
* fails closed instead of passing.
|
||||
*/
|
||||
/** Stage specifications — defines every pipeline stage. */
|
||||
export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
'00-intake': {
|
||||
number: '00',
|
||||
@@ -36,13 +27,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'research',
|
||||
gate: 'discovery-complete',
|
||||
promptFile: '00b-discovery.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'discovery-complete',
|
||||
reason: 'discovery completion is attested by an authority; no mechanical check exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
},
|
||||
'01-board': {
|
||||
number: '01',
|
||||
@@ -51,13 +36,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'review',
|
||||
gate: 'board-approval',
|
||||
promptFile: '01-board.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'board-approval',
|
||||
reason: 'board approval is a board/human decision; no mechanical gate exists',
|
||||
},
|
||||
],
|
||||
qualityGates: [{ type: 'ci-pipeline', command: 'board-approval (via board-tasks)' }],
|
||||
},
|
||||
'01b-brief-analyzer': {
|
||||
number: '01b',
|
||||
@@ -66,13 +45,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'research',
|
||||
gate: 'brief-analysis-complete',
|
||||
promptFile: '01-board.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'brief-analysis-complete',
|
||||
reason: 'brief analysis completion is attested by an authority; no mechanical check exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
},
|
||||
'02-planning-1': {
|
||||
number: '02',
|
||||
@@ -81,13 +54,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'research',
|
||||
gate: 'architecture-approval',
|
||||
promptFile: '02-planning-1-architecture.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'architecture-approval',
|
||||
reason: 'ADR approval requires authority sign-off; no mechanical check exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
},
|
||||
'03-planning-2': {
|
||||
number: '03',
|
||||
@@ -96,14 +63,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'research',
|
||||
gate: 'implementation-approval',
|
||||
promptFile: '03-planning-2-implementation.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'implementation-approval',
|
||||
reason:
|
||||
'implementation spec approval requires authority sign-off; no mechanical check exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
},
|
||||
'04-planning-3': {
|
||||
number: '04',
|
||||
@@ -112,14 +72,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'research',
|
||||
gate: 'decomposition-approval',
|
||||
promptFile: '04-planning-3-decomposition.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 'decomposition-approval',
|
||||
reason:
|
||||
'task decomposition approval requires authority sign-off; no mechanical check exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
},
|
||||
'05-coding': {
|
||||
number: '05',
|
||||
@@ -139,10 +92,9 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
promptFile: '06-review.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'provider',
|
||||
capability: 'reviewer',
|
||||
reason:
|
||||
'review verdicts require a wired reviewer provider; synthetic approvals are not permitted',
|
||||
type: 'ai-review',
|
||||
command:
|
||||
'echo \'{"summary":"review-pass","verdict":"approve","findings":[],"stats":{"blockers":0,"should_fix":0,"suggestions":0}}\'',
|
||||
},
|
||||
],
|
||||
},
|
||||
@@ -153,13 +105,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'coding',
|
||||
gate: 're-review',
|
||||
promptFile: '07-remediate.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'authority',
|
||||
capability: 're-review',
|
||||
reason: 'remediation re-review is an approval-based gate; no mechanical check exists',
|
||||
},
|
||||
],
|
||||
qualityGates: ['true'],
|
||||
},
|
||||
'08-test': {
|
||||
number: '08',
|
||||
@@ -177,13 +123,7 @@ export const STAGE_SPECS: Record<string, StageSpec> = {
|
||||
type: 'deploy',
|
||||
gate: 'deploy-verification',
|
||||
promptFile: '09-deploy.md',
|
||||
qualityGates: [
|
||||
{
|
||||
kind: 'provider',
|
||||
capability: 'ci-pipeline',
|
||||
reason: 'deploy verification requires a wired CI pipeline provider',
|
||||
},
|
||||
],
|
||||
qualityGates: [{ type: 'ci-pipeline', command: 'deploy-verification' }],
|
||||
},
|
||||
};
|
||||
|
||||
|
||||
@@ -1,46 +0,0 @@
|
||||
/**
|
||||
* Typed fail-closed capability errors (SDLC-D-035).
|
||||
*
|
||||
* A Forge run must fail closed when a required capability (executor, reviewer
|
||||
* provider, CI pipeline, authority sign-off) is missing. These typed errors
|
||||
* name the missing capability so callers can distinguish "not wired" from
|
||||
* ordinary execution failures.
|
||||
*/
|
||||
|
||||
/** Closed set of typed Forge capability error codes. */
|
||||
export const FORGE_ERROR_CODES = [
|
||||
'FORGE_NO_EXECUTOR',
|
||||
'FORGE_NO_REVIEWER',
|
||||
'FORGE_NO_CI_PIPELINE',
|
||||
'FORGE_NO_PROVIDER',
|
||||
'FORGE_AUTHORITY_REQUIRED',
|
||||
] as const;
|
||||
|
||||
export type ForgeErrorCode = (typeof FORGE_ERROR_CODES)[number];
|
||||
|
||||
/** Raised when a required capability is missing and the pipeline must fail closed. */
|
||||
export class ForgeCapabilityError extends Error {
|
||||
/** Typed error code from the closed FORGE_ERROR_CODES set. */
|
||||
readonly code: ForgeErrorCode;
|
||||
/** The missing capability, e.g. `task-executor`, `reviewer`, `board-approval`. */
|
||||
readonly capability: string;
|
||||
|
||||
constructor(code: ForgeErrorCode, capability: string, message: string) {
|
||||
super(message);
|
||||
this.name = 'ForgeCapabilityError';
|
||||
this.code = code;
|
||||
this.capability = capability;
|
||||
}
|
||||
}
|
||||
|
||||
/** Map a provider gate capability to its typed error code. */
|
||||
export function providerErrorCode(capability: string): ForgeErrorCode {
|
||||
switch (capability) {
|
||||
case 'reviewer':
|
||||
return 'FORGE_NO_REVIEWER';
|
||||
case 'ci-pipeline':
|
||||
return 'FORGE_NO_CI_PIPELINE';
|
||||
default:
|
||||
return 'FORGE_NO_PROVIDER';
|
||||
}
|
||||
}
|
||||
@@ -5,13 +5,6 @@ export type {
|
||||
StageSpec,
|
||||
BriefClass,
|
||||
ClassSource,
|
||||
ForgeOutcome,
|
||||
AuthorityGate,
|
||||
ProviderGate,
|
||||
ForgeGate,
|
||||
ForgeGateResult,
|
||||
ForgeTaskResult,
|
||||
RunMode,
|
||||
StageStatus,
|
||||
RunManifest,
|
||||
ForgeTaskStatus,
|
||||
@@ -88,24 +81,5 @@ export {
|
||||
getPipelineStatus,
|
||||
} from './pipeline-runner.js';
|
||||
|
||||
// Fail-closed errors and typed outcome model (SDLC-D-035)
|
||||
export { FORGE_ERROR_CODES, ForgeCapabilityError, providerErrorCode } from './errors.js';
|
||||
export type { ForgeErrorCode } from './errors.js';
|
||||
export {
|
||||
isSatisfyingOutcome,
|
||||
isCapabilityGate,
|
||||
isCommandGate,
|
||||
gateLabel,
|
||||
uniformGateResults,
|
||||
simulatedGateResults,
|
||||
waitingGateResults,
|
||||
blockedGateResults,
|
||||
evaluateStageGates,
|
||||
} from './outcomes.js';
|
||||
export type { StageEvaluation } from './outcomes.js';
|
||||
|
||||
// Simulated executor (explicit --simulate only)
|
||||
export { createSimulatedExecutor } from './simulated-executor.js';
|
||||
|
||||
// CLI
|
||||
export { registerForgeCommand } from './cli.js';
|
||||
|
||||
@@ -1,147 +0,0 @@
|
||||
import type { GateEntry } from '@mosaicstack/macp';
|
||||
|
||||
import type {
|
||||
AuthorityGate,
|
||||
ForgeGate,
|
||||
ForgeGateResult,
|
||||
ForgeOutcome,
|
||||
ForgeTaskResult,
|
||||
ProviderGate,
|
||||
} from './types.js';
|
||||
|
||||
/**
|
||||
* Gate and dependency satisfaction predicate (SDLC-D-035).
|
||||
*
|
||||
* ONLY a verified `passed` outcome satisfies. Every other member of the closed
|
||||
* outcome set — including `simulated` — is non-satisfying, so a simulated or
|
||||
* authority-blocked result can never be read as success-by-verification.
|
||||
*/
|
||||
export function isSatisfyingOutcome(outcome: ForgeOutcome): boolean {
|
||||
return outcome === 'passed';
|
||||
}
|
||||
|
||||
/** Whether a gate is an authority or provider gate (capability-based, command-less). */
|
||||
export function isCapabilityGate(gate: ForgeGate): gate is AuthorityGate | ProviderGate {
|
||||
if (typeof gate !== 'object' || gate === null) return false;
|
||||
const kind = (gate as Record<string, unknown>)['kind'];
|
||||
return kind === 'authority' || kind === 'provider';
|
||||
}
|
||||
|
||||
/** Whether a gate definition carries a real command a mechanical runner can execute. */
|
||||
export function isCommandGate(gate: ForgeGate): gate is string | GateEntry {
|
||||
if (typeof gate === 'string') {
|
||||
return gate.trim().length > 0;
|
||||
}
|
||||
if (isCapabilityGate(gate)) {
|
||||
// Authority and provider gates are satisfied by a capability, not a command.
|
||||
return false;
|
||||
}
|
||||
return typeof gate.command === 'string' && gate.command.trim().length > 0;
|
||||
}
|
||||
|
||||
/** Typed label identifying a gate in results and logs. */
|
||||
export function gateLabel(gate: ForgeGate): string {
|
||||
if (typeof gate === 'string') return gate;
|
||||
if (isCapabilityGate(gate)) return `${gate.kind}:${gate.capability}`;
|
||||
return gate.command || gate.type || 'unnamed-gate';
|
||||
}
|
||||
|
||||
/** Reason string stamped on every simulated gate result. */
|
||||
export const SIMULATED_GATE_REASON =
|
||||
'simulated execution (--simulate): gate was not evaluated by a real implementation';
|
||||
|
||||
/** Build typed gate results with a uniform outcome for a stage's declared gates. */
|
||||
export function uniformGateResults(
|
||||
gates: ForgeGate[],
|
||||
outcome: ForgeOutcome,
|
||||
reason: string,
|
||||
): ForgeGateResult[] {
|
||||
return gates.map((gate) => ({ gate: gateLabel(gate), outcome, reason }));
|
||||
}
|
||||
|
||||
/** Typed simulated gate results — used exclusively in `--simulate` runs. */
|
||||
export function simulatedGateResults(gates: ForgeGate[]): ForgeGateResult[] {
|
||||
return uniformGateResults(gates, 'simulated', SIMULATED_GATE_REASON);
|
||||
}
|
||||
|
||||
/** Typed waiting-for-authority gate results for approval-based stages. */
|
||||
export function waitingGateResults(gates: ForgeGate[], reason: string): ForgeGateResult[] {
|
||||
return uniformGateResults(gates, 'waiting-for-authority', reason);
|
||||
}
|
||||
|
||||
/** Typed blocked gate results for stages whose provider capability is not wired. */
|
||||
export function blockedGateResults(gates: ForgeGate[], reason: string): ForgeGateResult[] {
|
||||
return uniformGateResults(gates, 'blocked', reason);
|
||||
}
|
||||
|
||||
/** Outcome of evaluating a completed stage in normal mode. */
|
||||
export interface StageEvaluation {
|
||||
outcome: ForgeOutcome;
|
||||
reason: string;
|
||||
gateResults: ForgeGateResult[];
|
||||
}
|
||||
|
||||
/**
|
||||
* Evaluate a stage's declared gates against the executor's typed result.
|
||||
*
|
||||
* Fail-closed mapping:
|
||||
* - a `simulated` task or gate outcome in normal mode maps to `error`
|
||||
* - a missing gate result for a required command gate maps to `blocked`
|
||||
* - a non-passing task outcome propagates as the stage outcome
|
||||
* - only verified `passed` task and gate outcomes yield a `passed` stage
|
||||
*/
|
||||
export function evaluateStageGates(
|
||||
stageName: string,
|
||||
gates: ForgeGate[],
|
||||
result: ForgeTaskResult,
|
||||
): StageEvaluation {
|
||||
const gateResults = result.gate_results ?? [];
|
||||
|
||||
if (result.outcome === 'simulated') {
|
||||
return {
|
||||
outcome: 'error',
|
||||
reason: `executor reported a simulated outcome for stage '${stageName}' in normal mode — refusing to treat simulated results as verified`,
|
||||
gateResults,
|
||||
};
|
||||
}
|
||||
|
||||
if (!isSatisfyingOutcome(result.outcome)) {
|
||||
return {
|
||||
outcome: result.outcome,
|
||||
reason: `task outcome is '${result.outcome}': ${result.reason}`,
|
||||
gateResults,
|
||||
};
|
||||
}
|
||||
|
||||
for (const gate of gates) {
|
||||
// Authority and provider gates are pre-flighted before execution; they have
|
||||
// no mechanical result to verify here.
|
||||
if (!isCommandGate(gate)) continue;
|
||||
|
||||
const label = gateLabel(gate);
|
||||
const gateResult = gateResults.find((r) => r.gate === label);
|
||||
if (!gateResult) {
|
||||
return {
|
||||
outcome: 'blocked',
|
||||
reason: `no gate result was reported for required gate '${label}' (stage '${stageName}')`,
|
||||
gateResults,
|
||||
};
|
||||
}
|
||||
if (!isSatisfyingOutcome(gateResult.outcome)) {
|
||||
return {
|
||||
outcome: gateResult.outcome === 'simulated' ? 'error' : gateResult.outcome,
|
||||
reason: `gate '${label}' outcome is '${gateResult.outcome}': ${gateResult.reason}`,
|
||||
gateResults,
|
||||
};
|
||||
}
|
||||
}
|
||||
|
||||
return {
|
||||
outcome: 'passed',
|
||||
reason:
|
||||
gates.length === 0
|
||||
? "stage declares no gates; task outcome 'passed' accepted"
|
||||
: 'all declared gates verified passed',
|
||||
gateResults,
|
||||
};
|
||||
}
|
||||
@@ -1,33 +1,18 @@
|
||||
import fs from 'node:fs';
|
||||
import path from 'node:path';
|
||||
|
||||
import { STAGE_SEQUENCE, STAGE_SPECS } from './constants.js';
|
||||
import { STAGE_SEQUENCE } from './constants.js';
|
||||
import { determineBriefClass, stagesForClass } from './brief-classifier.js';
|
||||
import { ForgeCapabilityError, providerErrorCode } from './errors.js';
|
||||
import {
|
||||
blockedGateResults,
|
||||
evaluateStageGates,
|
||||
isCapabilityGate,
|
||||
simulatedGateResults,
|
||||
waitingGateResults,
|
||||
} from './outcomes.js';
|
||||
import { mapStageToTask } from './stage-adapter.js';
|
||||
import { createSimulatedExecutor } from './simulated-executor.js';
|
||||
import type {
|
||||
ForgeTask,
|
||||
ForgeTaskResult,
|
||||
PipelineOptions,
|
||||
PipelineResult,
|
||||
RunManifest,
|
||||
RunMode,
|
||||
StageStatus,
|
||||
TaskExecutor,
|
||||
} from './types.js';
|
||||
|
||||
/** Reason stamped on stages that complete under explicit simulation. */
|
||||
const SIMULATED_STAGE_REASON =
|
||||
'simulated execution (--simulate): stage was not executed by a real executor';
|
||||
|
||||
/**
|
||||
* Generate a timestamp-based run ID.
|
||||
*/
|
||||
@@ -62,7 +47,6 @@ function createManifest(opts: {
|
||||
briefClass: RunManifest['briefClass'];
|
||||
classSource: RunManifest['classSource'];
|
||||
forceBoard: boolean;
|
||||
mode: RunMode;
|
||||
runDir: string;
|
||||
}): RunManifest {
|
||||
const ts = nowISO();
|
||||
@@ -73,7 +57,6 @@ function createManifest(opts: {
|
||||
briefClass: opts.briefClass,
|
||||
classSource: opts.classSource,
|
||||
forceBoard: opts.forceBoard,
|
||||
mode: opts.mode,
|
||||
createdAt: ts,
|
||||
updatedAt: ts,
|
||||
currentStage: '',
|
||||
@@ -125,199 +108,20 @@ export function selectStages(stages?: string[], skipTo?: string): string[] {
|
||||
return selected.slice(skipIndex);
|
||||
}
|
||||
|
||||
/**
|
||||
* Fail closed when the required executor capability is missing (SDLC-D-035).
|
||||
*/
|
||||
function requireExecutor(executor: TaskExecutor | undefined, simulate: boolean): TaskExecutor {
|
||||
if (executor) return executor;
|
||||
if (simulate) return createSimulatedExecutor({ log: false });
|
||||
throw new ForgeCapabilityError(
|
||||
'FORGE_NO_EXECUTOR',
|
||||
'task-executor',
|
||||
'no task executor is wired; refusing to run the pipeline with a stub executor (fail closed). ' +
|
||||
'Pass --simulate to opt into explicitly simulated execution.',
|
||||
);
|
||||
}
|
||||
|
||||
/**
|
||||
* Pre-flight a stage's gates in normal mode (fail closed, SDLC-D-035).
|
||||
*
|
||||
* - authority gates: record a typed `waiting-for-authority` stage result and
|
||||
* raise FORGE_AUTHORITY_REQUIRED — approval-based gates never pass vacuously.
|
||||
* - provider gates: record a typed `blocked` stage result and raise the typed
|
||||
* capability error for the missing provider.
|
||||
*
|
||||
* Returns the stage status to record when the pre-flight blocks, or undefined
|
||||
* when the stage may proceed.
|
||||
*/
|
||||
function preflightStageGates(
|
||||
stageName: string,
|
||||
manifest: RunManifest,
|
||||
): { status: StageStatus; error: ForgeCapabilityError } | undefined {
|
||||
const spec = STAGE_SPECS[stageName];
|
||||
if (!spec) throw new Error(`Unknown Forge stage: ${stageName}`);
|
||||
|
||||
for (const gate of spec.qualityGates) {
|
||||
if (!isCapabilityGate(gate)) continue;
|
||||
|
||||
const startedAt = manifest.stages[stageName]?.startedAt;
|
||||
const completedAt = nowISO();
|
||||
|
||||
if (gate.kind === 'authority') {
|
||||
const reason = `gate '${gate.capability}' requires authority sign-off; no mechanical implementation exists (${gate.reason})`;
|
||||
return {
|
||||
status: {
|
||||
status: 'waiting-for-authority',
|
||||
reason,
|
||||
startedAt,
|
||||
completedAt,
|
||||
gateResults: waitingGateResults(spec.qualityGates, reason),
|
||||
},
|
||||
error: new ForgeCapabilityError(
|
||||
'FORGE_AUTHORITY_REQUIRED',
|
||||
gate.capability,
|
||||
`stage '${stageName}' is blocked on authority gate '${gate.capability}': ${gate.reason}. ` +
|
||||
'The pipeline fails closed instead of passing vacuously. Record the approval out-of-band ' +
|
||||
'or run with --simulate for explicitly simulated execution.',
|
||||
),
|
||||
};
|
||||
}
|
||||
|
||||
const reason = `gate '${gate.capability}' requires provider '${gate.capability}' and none is wired (${gate.reason})`;
|
||||
return {
|
||||
status: {
|
||||
status: 'blocked',
|
||||
reason,
|
||||
startedAt,
|
||||
completedAt,
|
||||
gateResults: blockedGateResults(spec.qualityGates, reason),
|
||||
},
|
||||
error: new ForgeCapabilityError(
|
||||
providerErrorCode(gate.capability),
|
||||
gate.capability,
|
||||
`stage '${stageName}' requires provider '${gate.capability}' which is not wired: ${gate.reason}. ` +
|
||||
'The pipeline fails closed instead of passing vacuously.',
|
||||
),
|
||||
};
|
||||
}
|
||||
|
||||
return undefined;
|
||||
}
|
||||
|
||||
/**
|
||||
* Execute the given stage tasks sequentially, updating the manifest.
|
||||
*
|
||||
* Normal mode requires a real executor and evaluates every declared command
|
||||
* gate through the typed outcome model; any non-verified result fails closed.
|
||||
* Simulate mode types every stage and gate result as `simulated`.
|
||||
*/
|
||||
async function executeStages(opts: {
|
||||
manifest: RunManifest;
|
||||
runDir: string;
|
||||
tasks: ForgeTask[];
|
||||
stageNames: string[];
|
||||
executor: TaskExecutor;
|
||||
simulate: boolean;
|
||||
}): Promise<void> {
|
||||
const { manifest, runDir, tasks, stageNames, executor, simulate } = opts;
|
||||
|
||||
for (let i = 0; i < tasks.length; i++) {
|
||||
const task = tasks[i]!;
|
||||
const stageName = stageNames[i]!;
|
||||
const spec = STAGE_SPECS[stageName];
|
||||
if (!spec) throw new Error(`Unknown Forge stage: ${stageName}`);
|
||||
|
||||
// Update manifest: stage in progress
|
||||
manifest.currentStage = stageName;
|
||||
manifest.stages[stageName] = {
|
||||
status: 'in_progress',
|
||||
startedAt: nowISO(),
|
||||
};
|
||||
saveManifest(runDir, manifest);
|
||||
|
||||
// Fail-closed pre-flight (normal mode only): authority/provider gates have
|
||||
// no mechanical implementation and must never pass vacuously.
|
||||
if (!simulate) {
|
||||
const blocked = preflightStageGates(stageName, manifest);
|
||||
if (blocked) {
|
||||
manifest.stages[stageName] = blocked.status;
|
||||
manifest.status =
|
||||
blocked.status.status === 'waiting-for-authority' ? 'waiting-for-authority' : 'failed';
|
||||
saveManifest(runDir, manifest);
|
||||
throw blocked.error;
|
||||
}
|
||||
}
|
||||
|
||||
let result: ForgeTaskResult;
|
||||
try {
|
||||
await executor.submitTask(task);
|
||||
result = await executor.waitForCompletion(task.id, task.timeoutSeconds * 1000);
|
||||
} catch (error) {
|
||||
// Process errors (including timeouts) map to the fail-closed `error` outcome.
|
||||
const reason = error instanceof Error ? error.message : String(error);
|
||||
manifest.stages[stageName] = {
|
||||
status: 'error',
|
||||
reason: `executor error: ${reason}`,
|
||||
startedAt: manifest.stages[stageName]?.startedAt,
|
||||
completedAt: nowISO(),
|
||||
gateResults: [],
|
||||
};
|
||||
manifest.status = 'failed';
|
||||
saveManifest(runDir, manifest);
|
||||
throw error instanceof Error ? error : new Error(reason);
|
||||
}
|
||||
|
||||
if (simulate) {
|
||||
manifest.stages[stageName] = {
|
||||
status: 'simulated',
|
||||
reason: SIMULATED_STAGE_REASON,
|
||||
startedAt: manifest.stages[stageName]?.startedAt,
|
||||
completedAt: nowISO(),
|
||||
gateResults: simulatedGateResults(spec.qualityGates),
|
||||
};
|
||||
saveManifest(runDir, manifest);
|
||||
continue;
|
||||
}
|
||||
|
||||
const evaluation = evaluateStageGates(stageName, spec.qualityGates, result);
|
||||
manifest.stages[stageName] = {
|
||||
status: evaluation.outcome,
|
||||
reason: evaluation.reason,
|
||||
startedAt: manifest.stages[stageName]?.startedAt,
|
||||
completedAt: nowISO(),
|
||||
gateResults: evaluation.gateResults,
|
||||
};
|
||||
|
||||
if (evaluation.outcome !== 'passed') {
|
||||
manifest.status =
|
||||
evaluation.outcome === 'waiting-for-authority' ? 'waiting-for-authority' : 'failed';
|
||||
saveManifest(runDir, manifest);
|
||||
throw new Error(`Stage ${stageName} ${evaluation.outcome}: ${evaluation.reason}`);
|
||||
}
|
||||
|
||||
saveManifest(runDir, manifest);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Run the Forge pipeline.
|
||||
*
|
||||
* 1. Fail closed unless a real executor is wired or simulation is explicit
|
||||
* 2. Classify the brief
|
||||
* 3. Generate a run ID and create run directory
|
||||
* 4. Map stages to tasks and submit to TaskExecutor
|
||||
* 5. Track manifest with typed stage outcomes
|
||||
* 6. Return pipeline result
|
||||
* 1. Classify the brief
|
||||
* 2. Generate a run ID and create run directory
|
||||
* 3. Map stages to tasks and submit to TaskExecutor
|
||||
* 4. Track manifest with stage statuses
|
||||
* 5. Return pipeline result
|
||||
*/
|
||||
export async function runPipeline(
|
||||
briefPath: string,
|
||||
projectRoot: string,
|
||||
options: PipelineOptions,
|
||||
): Promise<PipelineResult> {
|
||||
const simulate = options.simulate ?? false;
|
||||
const executor = requireExecutor(options.executor, simulate);
|
||||
const mode: RunMode = simulate ? 'simulated' : 'normal';
|
||||
|
||||
const resolvedRoot = path.resolve(projectRoot);
|
||||
const resolvedBrief = path.resolve(briefPath);
|
||||
const briefContent = fs.readFileSync(resolvedBrief, 'utf-8');
|
||||
@@ -342,7 +146,6 @@ export async function runPipeline(
|
||||
briefClass,
|
||||
classSource,
|
||||
forceBoard: options.forceBoard ?? false,
|
||||
mode,
|
||||
runDir,
|
||||
});
|
||||
|
||||
@@ -369,10 +172,54 @@ export async function runPipeline(
|
||||
}
|
||||
|
||||
// Execute stages
|
||||
await executeStages({ manifest, runDir, tasks, stageNames: selectedStages, executor, simulate });
|
||||
const { executor } = options;
|
||||
for (let i = 0; i < tasks.length; i++) {
|
||||
const task = tasks[i]!;
|
||||
const stageName = selectedStages[i]!;
|
||||
|
||||
// All stages reached a terminal state for this mode
|
||||
manifest.status = simulate ? 'simulated' : 'completed';
|
||||
// Update manifest: stage in progress
|
||||
manifest.currentStage = stageName;
|
||||
manifest.stages[stageName] = {
|
||||
status: 'in_progress',
|
||||
startedAt: nowISO(),
|
||||
};
|
||||
saveManifest(runDir, manifest);
|
||||
|
||||
try {
|
||||
await executor.submitTask(task);
|
||||
const result = await executor.waitForCompletion(task.id, task.timeoutSeconds * 1000);
|
||||
|
||||
// Update manifest: stage completed or failed
|
||||
const stageStatus: StageStatus = {
|
||||
status: result.status === 'completed' ? 'passed' : 'failed',
|
||||
startedAt: manifest.stages[stageName]!.startedAt,
|
||||
completedAt: nowISO(),
|
||||
};
|
||||
manifest.stages[stageName] = stageStatus;
|
||||
|
||||
if (result.status !== 'completed') {
|
||||
manifest.status = 'failed';
|
||||
saveManifest(runDir, manifest);
|
||||
throw new Error(`Stage ${stageName} failed with status: ${result.status}`);
|
||||
}
|
||||
|
||||
saveManifest(runDir, manifest);
|
||||
} catch (error) {
|
||||
if (!manifest.stages[stageName]?.completedAt) {
|
||||
manifest.stages[stageName] = {
|
||||
status: 'failed',
|
||||
startedAt: manifest.stages[stageName]?.startedAt,
|
||||
completedAt: nowISO(),
|
||||
};
|
||||
}
|
||||
manifest.status = 'failed';
|
||||
saveManifest(runDir, manifest);
|
||||
throw error;
|
||||
}
|
||||
}
|
||||
|
||||
// All stages passed
|
||||
manifest.status = 'completed';
|
||||
saveManifest(runDir, manifest);
|
||||
|
||||
return {
|
||||
@@ -387,30 +234,22 @@ export async function runPipeline(
|
||||
}
|
||||
|
||||
/**
|
||||
* Resume a pipeline from the last non-passed stage.
|
||||
* Resume a pipeline from the last incomplete stage.
|
||||
*/
|
||||
export async function resumePipeline(
|
||||
runDir: string,
|
||||
executor?: TaskExecutor,
|
||||
options?: { simulate?: boolean },
|
||||
executor: TaskExecutor,
|
||||
): Promise<PipelineResult> {
|
||||
const simulate = options?.simulate ?? false;
|
||||
const wiredExecutor = requireExecutor(executor, simulate);
|
||||
const mode: RunMode = simulate ? 'simulated' : 'normal';
|
||||
|
||||
const manifest = loadManifest(runDir);
|
||||
const resolvedRoot = path.dirname(path.dirname(path.dirname(runDir))); // .forge/runs/{id} → project root
|
||||
|
||||
const briefContent = fs.readFileSync(manifest.brief, 'utf-8');
|
||||
const allStages = stagesForClass(manifest.briefClass, manifest.forceBoard);
|
||||
|
||||
manifest.mode = mode;
|
||||
|
||||
// Find first non-satisfying stage (only a verified `passed` counts as done;
|
||||
// simulated and waiting-for-authority stages are re-run).
|
||||
// Find first non-passed stage
|
||||
const resumeFrom = allStages.find((s) => manifest.stages[s]?.status !== 'passed');
|
||||
if (!resumeFrom) {
|
||||
manifest.status = mode === 'simulated' ? 'simulated' : 'completed';
|
||||
manifest.status = 'completed';
|
||||
saveManifest(runDir, manifest);
|
||||
return {
|
||||
runId: manifest.runId,
|
||||
@@ -445,16 +284,49 @@ export async function resumePipeline(
|
||||
tasks.push(task);
|
||||
}
|
||||
|
||||
await executeStages({
|
||||
manifest,
|
||||
runDir,
|
||||
tasks,
|
||||
stageNames: remainingStages,
|
||||
executor: wiredExecutor,
|
||||
simulate,
|
||||
});
|
||||
for (let i = 0; i < tasks.length; i++) {
|
||||
const task = tasks[i]!;
|
||||
const stageName = remainingStages[i]!;
|
||||
|
||||
manifest.status = simulate ? 'simulated' : 'completed';
|
||||
manifest.currentStage = stageName;
|
||||
manifest.stages[stageName] = {
|
||||
status: 'in_progress',
|
||||
startedAt: nowISO(),
|
||||
};
|
||||
saveManifest(runDir, manifest);
|
||||
|
||||
try {
|
||||
await executor.submitTask(task);
|
||||
const result = await executor.waitForCompletion(task.id, task.timeoutSeconds * 1000);
|
||||
|
||||
manifest.stages[stageName] = {
|
||||
status: result.status === 'completed' ? 'passed' : 'failed',
|
||||
startedAt: manifest.stages[stageName]!.startedAt,
|
||||
completedAt: nowISO(),
|
||||
};
|
||||
|
||||
if (result.status !== 'completed') {
|
||||
manifest.status = 'failed';
|
||||
saveManifest(runDir, manifest);
|
||||
throw new Error(`Stage ${stageName} failed with status: ${result.status}`);
|
||||
}
|
||||
|
||||
saveManifest(runDir, manifest);
|
||||
} catch (error) {
|
||||
if (!manifest.stages[stageName]?.completedAt) {
|
||||
manifest.stages[stageName] = {
|
||||
status: 'failed',
|
||||
startedAt: manifest.stages[stageName]?.startedAt,
|
||||
completedAt: nowISO(),
|
||||
};
|
||||
}
|
||||
manifest.status = 'failed';
|
||||
saveManifest(runDir, manifest);
|
||||
throw error;
|
||||
}
|
||||
}
|
||||
|
||||
manifest.status = 'completed';
|
||||
saveManifest(runDir, manifest);
|
||||
|
||||
return {
|
||||
|
||||
@@ -1,32 +0,0 @@
|
||||
import type { ForgeTask, ForgeTaskResult, TaskExecutor } from './types.js';
|
||||
|
||||
/**
|
||||
* Simulated executor — used ONLY when the caller explicitly passes --simulate.
|
||||
*
|
||||
* It submits no real work and returns typed `simulated` results so a simulated
|
||||
* run can never be confused with a verified one. In normal mode (no --simulate)
|
||||
* the CLI refuses to run at all with FORGE_NO_EXECUTOR instead of wiring this
|
||||
* stub (fail closed, SDLC-D-035).
|
||||
*/
|
||||
export function createSimulatedExecutor(options?: { log?: boolean }): TaskExecutor {
|
||||
const log = options?.log ?? true;
|
||||
return {
|
||||
async submitTask(task: ForgeTask) {
|
||||
if (log) console.log(` [forge:simulated] stage submitted: ${task.id} (${task.title})`);
|
||||
},
|
||||
async waitForCompletion(taskId: string): Promise<ForgeTaskResult> {
|
||||
if (log) console.log(` [forge:simulated] stage complete: ${taskId}`);
|
||||
return {
|
||||
task_id: taskId,
|
||||
outcome: 'simulated',
|
||||
reason: 'no executor wired; simulated execution requested via --simulate',
|
||||
completed_at: new Date().toISOString(),
|
||||
exit_code: 0,
|
||||
gate_results: [],
|
||||
};
|
||||
},
|
||||
async getTaskStatus() {
|
||||
return 'completed' as const;
|
||||
},
|
||||
};
|
||||
}
|
||||
@@ -1,4 +1,4 @@
|
||||
import type { GateEntry } from '@mosaicstack/macp';
|
||||
import type { GateEntry, TaskResult } from '@mosaicstack/macp';
|
||||
|
||||
/** Stage dispatch mode. */
|
||||
export type StageDispatch = 'exec' | 'yolo' | 'pi';
|
||||
@@ -6,58 +6,6 @@ export type StageDispatch = 'exec' | 'yolo' | 'pi';
|
||||
/** Stage type — determines agent selection and gate requirements. */
|
||||
export type StageType = 'research' | 'review' | 'coding' | 'deploy';
|
||||
|
||||
/**
|
||||
* Typed outcome for every gate and stage evaluation — closed set (SDLC-D-035).
|
||||
*
|
||||
* Only `passed` means "verified by a real implementation". `simulated` is
|
||||
* produced exclusively in explicit `--simulate` runs and is never satisfying.
|
||||
*/
|
||||
export type ForgeOutcome =
|
||||
| 'passed'
|
||||
| 'failed'
|
||||
| 'blocked'
|
||||
| 'error'
|
||||
| 'waiting-for-authority'
|
||||
| 'simulated'
|
||||
| 'not-applicable';
|
||||
|
||||
/** A gate that requires authority (human/board) sign-off; no mechanical command can satisfy it. */
|
||||
export interface AuthorityGate {
|
||||
kind: 'authority';
|
||||
capability: string;
|
||||
reason: string;
|
||||
}
|
||||
|
||||
/** A gate that requires a wired provider (e.g. an AI reviewer, CI pipeline) to evaluate. */
|
||||
export interface ProviderGate {
|
||||
kind: 'provider';
|
||||
capability: string;
|
||||
reason: string;
|
||||
}
|
||||
|
||||
/** Forge quality gate: a real command, an authority sign-off, or a provider-backed check. */
|
||||
export type ForgeGate = string | GateEntry | AuthorityGate | ProviderGate;
|
||||
|
||||
/** Typed result of evaluating a single quality gate. */
|
||||
export interface ForgeGateResult {
|
||||
gate: string;
|
||||
outcome: ForgeOutcome;
|
||||
reason: string;
|
||||
exitCode?: number;
|
||||
output?: string;
|
||||
timedOut?: boolean;
|
||||
}
|
||||
|
||||
/** Typed result of a task/stage execution returned by a TaskExecutor. */
|
||||
export interface ForgeTaskResult {
|
||||
task_id: string;
|
||||
outcome: ForgeOutcome;
|
||||
reason: string;
|
||||
completed_at: string;
|
||||
exit_code: number;
|
||||
gate_results: ForgeGateResult[];
|
||||
}
|
||||
|
||||
/** Stage specification — defines a single pipeline stage. */
|
||||
export interface StageSpec {
|
||||
number: string;
|
||||
@@ -66,7 +14,7 @@ export interface StageSpec {
|
||||
type: StageType;
|
||||
gate: string;
|
||||
promptFile: string;
|
||||
qualityGates: ForgeGate[];
|
||||
qualityGates: (string | GateEntry)[];
|
||||
}
|
||||
|
||||
/** Brief classification. */
|
||||
@@ -77,18 +25,11 @@ export type ClassSource = 'cli' | 'frontmatter' | 'auto';
|
||||
|
||||
/** Per-stage status within a run manifest. */
|
||||
export interface StageStatus {
|
||||
status: 'pending' | 'in_progress' | ForgeOutcome;
|
||||
/** Why the stage reached its current (terminal) outcome, when applicable. */
|
||||
reason?: string;
|
||||
status: 'pending' | 'in_progress' | 'passed' | 'failed';
|
||||
startedAt?: string;
|
||||
completedAt?: string;
|
||||
/** Typed per-gate results recorded alongside the stage outcome. */
|
||||
gateResults?: ForgeGateResult[];
|
||||
}
|
||||
|
||||
/** Execution mode of a run. */
|
||||
export type RunMode = 'normal' | 'simulated';
|
||||
|
||||
/** Run manifest — persisted to disk as manifest.json. */
|
||||
export interface RunManifest {
|
||||
runId: string;
|
||||
@@ -97,23 +38,10 @@ export interface RunManifest {
|
||||
briefClass: BriefClass;
|
||||
classSource: ClassSource;
|
||||
forceBoard: boolean;
|
||||
/**
|
||||
* Execution mode. `simulated` runs stub execution; their results are typed
|
||||
* `simulated` and must never be read as verified success. Optional because
|
||||
* manifests written before this field existed default to `normal`.
|
||||
*/
|
||||
mode?: RunMode;
|
||||
createdAt: string;
|
||||
updatedAt: string;
|
||||
currentStage: string;
|
||||
status:
|
||||
| 'in_progress'
|
||||
| 'completed'
|
||||
| 'failed'
|
||||
| 'interrupted'
|
||||
| 'rejected'
|
||||
| 'simulated'
|
||||
| 'waiting-for-authority';
|
||||
status: 'in_progress' | 'completed' | 'failed' | 'interrupted' | 'rejected';
|
||||
stages: Record<string, StageStatus>;
|
||||
}
|
||||
|
||||
@@ -137,7 +65,7 @@ export interface ForgeTask {
|
||||
briefPath: string;
|
||||
resultPath: string;
|
||||
timeoutSeconds: number;
|
||||
qualityGates: ForgeGate[];
|
||||
qualityGates: (string | GateEntry)[];
|
||||
worktree?: string;
|
||||
command?: string;
|
||||
dependsOn?: string[];
|
||||
@@ -148,7 +76,7 @@ export interface ForgeTask {
|
||||
/** Abstract task executor — decouples from packages/coord. */
|
||||
export interface TaskExecutor {
|
||||
submitTask(task: ForgeTask): Promise<void>;
|
||||
waitForCompletion(taskId: string, timeoutMs: number): Promise<ForgeTaskResult>;
|
||||
waitForCompletion(taskId: string, timeoutMs: number): Promise<TaskResult>;
|
||||
getTaskStatus(taskId: string): Promise<ForgeTaskStatus>;
|
||||
}
|
||||
|
||||
@@ -194,16 +122,7 @@ export interface PipelineOptions {
|
||||
stages?: string[];
|
||||
skipTo?: string;
|
||||
dryRun?: boolean;
|
||||
/**
|
||||
* Real task executor. Required in normal mode: the pipeline fails closed
|
||||
* with FORGE_NO_EXECUTOR when it is absent.
|
||||
*/
|
||||
executor?: TaskExecutor;
|
||||
/**
|
||||
* Explicit opt-in to simulated execution. Every stage and gate result is
|
||||
* typed `simulated` and is never satisfying.
|
||||
*/
|
||||
simulate?: boolean;
|
||||
executor: TaskExecutor;
|
||||
}
|
||||
|
||||
/** Pipeline run result. */
|
||||
|
||||
@@ -39,20 +39,3 @@ packages/mosaic/framework/tools/tmux/test-send-message-verdict.sh | requires rea
|
||||
# recorded judgement. These lines ARE that judgement, signed.)
|
||||
packages/mosaic/framework/tools/orchestrator/smoke-test.sh | behavior smoke checks for coord continue/run workflows, run manually by orchestrator seats; unmeasured in CI; #1017 burndown
|
||||
packages/mosaic/framework/tools/wake/validate-973/microtest-wake-assert.sh | #973 instrument self-test, run as a precondition of the validate-973 evidence procedure rather than as a standing CI suite; #1017 burndown candidate
|
||||
|
||||
# --- tools/fleet: precondition is unsatisfiable in the CI image (#1271) ---
|
||||
# Signed by fred (sb-it-1-dt, 2026-08-16) at origin/next 476db12.
|
||||
# This suite asserts the launcher's behaviour when `mosaic` and `pi` are MISSING.
|
||||
# It shims fakes into $FAKE_BIN, but the constructed PANE_PATH always ends in the
|
||||
# real system path, so on a host that installs those binaries the missing-binary
|
||||
# cases cannot be measured at all. The suite's own guard (line 103) says so and
|
||||
# fails rather than reporting a pass it cannot back. That guard is correct.
|
||||
# The error was wiring the suite into CI: #1017 (c56483eb) enumerated it and
|
||||
# dropped this exclusion, and the CI image provides `pi` in the system path, so
|
||||
# it has failed on every pipeline since. Measured 2026-08-16 across pipelines
|
||||
# 2444 (#1256), 2438 (#1240) and 2441 (#1017-quality): exactly one FAIL line in
|
||||
# each full log, identical, this assertion; control `zzz-not-present-zzz` -> 0.
|
||||
# Burn-down and the full measurement are tracked in #1271; unwired by PR #1270.
|
||||
# Because test:framework-shell is one && chain and this sat at position 44 of 48,
|
||||
# the four suites after it had not run at all since the merge.
|
||||
packages/mosaic/framework/tools/fleet/test-start-agent-session.sh | precondition unsatisfiable in the CI image: asserts missing-binary behaviour, but PANE_PATH always ends in the system path and the image provides `pi` there; guard at line 103 fails by design rather than passing unmeasured. Burn down by controlling the tail of PANE_PATH inside the test. NOT by removing `pi` from the image: the CI image installs @earendil-works/[email protected] deliberately (measured in pipeline 2444's test-step log), and other suites depend on that pin. Burn-down tracked in #1271
|
||||
|
||||
@@ -97,13 +97,34 @@ printf '%s' "$MSG" | "${tmux_cmd[@]}" load-buffer -b "$BUF" -
|
||||
# would otherwise accumulate forever.
|
||||
sleep 0.5
|
||||
|
||||
# 2) Submit, then POSITIVELY confirm submission; flush with another Enter if it is
|
||||
# still a draft. Success requires positive evidence — the queued banner, OR the
|
||||
# REPL input box located AND clear of our message tail. The historical bug was
|
||||
# treating ABSENCE of a draft as delivery: if the prompt glyph was never matched
|
||||
# (wrong pane / prompt-glyph drift), an unsubmitted message read as "delivered"
|
||||
# and worker->lead relays stalled silently. We now default to UNCONFIRMED and only
|
||||
# upgrade to delivered on positive evidence; anything we cannot confirm fails loud.
|
||||
# 2) Submit, then POSITIVELY confirm submission by DRAFT TRANSITION, not by prompt
|
||||
# glyph. The historical bug was treating ABSENCE of a draft as delivery; the
|
||||
# 2026-08 fix over-corrected to glyph inference (grep '❯|^>|│ >'), which locates
|
||||
# only Claude Code's box and false-NEGATIVES every glyphless REPL (pi renders a
|
||||
# U+2500 rule, no glyph) — a delivered message reported "UNDELIVERED", driving a
|
||||
# retry that duplicates it. Runtime-agnostic evidence: our message tail sits on
|
||||
# the INPUT line (located by the cursor row, not a glyph) BEFORE Enter, and has
|
||||
# LEFT it AFTER — that transition is positive proof of submission and needs no
|
||||
# glyph. Absence alone still never means delivered: if we never saw our draft on
|
||||
# the input line we stay UNCONFIRMED (wrong/dead pane), and a draft that never
|
||||
# leaves the input line stays a DRAFT (exit 2), preserving both historical guards.
|
||||
_cursor_line() { # echo the pane's current input (cursor) line, glyph-free
|
||||
local cy line
|
||||
cy=$("${tmux_cmd[@]}" display-message -p -t "$EFFECTIVE_TARGET" -F '#{cursor_y}' 2>/dev/null) || return 1
|
||||
[ -n "$cy" ] || return 1
|
||||
"${tmux_cmd[@]}" capture-pane -t "$EFFECTIVE_TARGET" -p 2>/dev/null | sed -n "$((cy + 1))p"
|
||||
}
|
||||
_draft_on_input() { # true iff our message tail is sitting on the input line now
|
||||
[ -n "$snippet" ] || return 1
|
||||
printf '%s' "$(_cursor_line)" | grep -qF "$snippet"
|
||||
}
|
||||
|
||||
# Baseline: after the paste, our draft must be on the input line. This is positive
|
||||
# proof we are on the right pane and the paste landed — the anchor the transition
|
||||
# check measures against.
|
||||
saw_draft=0
|
||||
_draft_on_input && saw_draft=1
|
||||
|
||||
status="unconfirmed"
|
||||
for attempt in $(seq 1 $((RETRIES + 1))); do
|
||||
"${tmux_cmd[@]}" send-keys -t "$EFFECTIVE_TARGET" Enter
|
||||
@@ -113,20 +134,26 @@ for attempt in $(seq 1 $((RETRIES + 1))); do
|
||||
if printf '%s' "$pane" | grep -qF "$QUEUED_RE"; then
|
||||
status="queued"; break
|
||||
fi
|
||||
# Locate the REPL input box (prompt glyph). If we cannot see it, we have NO
|
||||
# evidence of submission state — stay UNCONFIRMED and retry; never infer delivery.
|
||||
# POSITIVE draft evidence from a located prompt box, when one exists. This is the
|
||||
# cursor-row check's blind spot: a pane in COOKED mode (a plain shell whose
|
||||
# foreground process never reads stdin) echoes our paste via the kernel line
|
||||
# discipline and moves the cursor off it on Enter, which is indistinguishable from
|
||||
# a real submit by cursor row alone. If a prompt box IS locatable and still carries
|
||||
# our tail, that is affirmative proof the message was not consumed. Absence of a
|
||||
# glyph is still never used for anything — that inference is the original E7 bug.
|
||||
promptline=$(printf '%s' "$pane" | grep -E '❯|^>|│ >' | tail -1)
|
||||
if [ -z "$promptline" ]; then
|
||||
status="unconfirmed"; continue
|
||||
fi
|
||||
# Input box located AND still carrying our tail => unsubmitted draft. Flush + retry.
|
||||
# (Submitted messages scroll up into history; a draft stays on the ❯ line.)
|
||||
if [ -n "$snippet" ] && printf '%s' "$promptline" | grep -qF "$snippet"; then
|
||||
if [ -n "$promptline" ] && [ -n "$snippet" ] && printf '%s' "$promptline" | grep -qF "$snippet"; then
|
||||
status="draft"; continue
|
||||
fi
|
||||
# Input box located AND clear of our tail => positively submitted. This is the
|
||||
# only path to success besides the queued banner.
|
||||
status="delivered"; break
|
||||
if [ "$saw_draft" = 1 ]; then
|
||||
if _draft_on_input; then
|
||||
status="draft"; continue # still on the input line => not submitted; flush + retry
|
||||
fi
|
||||
status="delivered"; break # left the input line => positively submitted
|
||||
fi
|
||||
# No confirmed baseline yet: try to (re)acquire it; never infer delivery from absence.
|
||||
if _draft_on_input; then saw_draft=1; status="draft"; continue; fi
|
||||
status="unconfirmed"; continue
|
||||
done
|
||||
|
||||
[ "$VERBOSE" = 1 ] && { echo "--- pane tail ($TARGET) ---"; printf '%s\n' "$pane" | tail -4; echo "---"; }
|
||||
|
||||
@@ -0,0 +1,97 @@
|
||||
#!/usr/bin/env bash
|
||||
# Red-first regression test for E7 (#1017 task 2): the confirm-check must bind
|
||||
# "delivered" to WHETHER THE MESSAGE WAS SUBMITTED, not to which runtime's prompt
|
||||
# glyph is present. A pi seat renders a U+2500 rule input box with no ❯/^>/│ >
|
||||
# glyph; send-message.sh:118 locates the box only by glyph, so a genuinely
|
||||
# delivered message on a glyphless REPL falsely reports exit 2 "may be UNDELIVERED",
|
||||
# and the operator's rc=2-driven retry duplicates it.
|
||||
#
|
||||
# Parameterized on $SEND: RED against the shipping blob (B and D fail), GREEN
|
||||
# against a candidate patch. No pi; no fake HOME; hermetic throwaway socket.
|
||||
#
|
||||
# Submission counting is EXACT and terminal-echo-independent: the fixture message
|
||||
# is `echo <tok> >>SINK`; each real submission appends one line. wc -l SINK ==
|
||||
# number of times the REPL actually executed the send. This does not depend on how
|
||||
# many times the marker string is painted on screen.
|
||||
set -u
|
||||
SEND="${SEND:?set SEND=/path/to/send-message.sh}"
|
||||
SOCKET="glyphagnostic-$$"
|
||||
TMP="$(mktemp -d)"
|
||||
tmux() { command tmux -L "$SOCKET" "$@"; }
|
||||
cleanup() { command tmux -L "$SOCKET" kill-server 2>/dev/null; rm -rf "$TMP"; }
|
||||
trap cleanup EXIT
|
||||
pass=0; fail=0
|
||||
ok() { printf 'ok %s\n' "$1"; pass=$((pass+1)); }
|
||||
no() { printf 'FAIL %s -- %s\n' "$1" "$2"; fail=$((fail+1)); }
|
||||
|
||||
mk() { tmux new-session -d -s "$1" -x 120 -y 40 -c "$TMP" "PS1='$2' exec bash --noprofile --norc -i"; sleep 0.5; }
|
||||
subs() { [ -f "$1" ] && wc -l <"$1" | tr -d ' ' || echo 0; } # exact submission count
|
||||
|
||||
echo "SEND=$SEND tmux $(command tmux -V | awk '{print $2}')"
|
||||
|
||||
# --- A (control): glyph box (❯) that submits => exit 0, exactly one submission.
|
||||
mk ctl '❯ '
|
||||
SINK="$TMP/sink.ctl"
|
||||
out=$("$SEND" -L "$SOCKET" -t ctl -m "echo x >>'$SINK'" 2>"$TMP/e.ctl"); rc=$?; sleep 0.4
|
||||
if [ "$rc" = 0 ] && [ "$(subs "$SINK")" = 1 ]; then
|
||||
ok "control: ❯-box submits => exit 0, exactly one submission"
|
||||
else no "control: ❯-box submits => exit 0, one submission" "rc=$rc subs=$(subs "$SINK") err=[$(cat "$TMP/e.ctl")]"; fi
|
||||
|
||||
# --- B (THE false-rc regression): glyphless U+2500 box that SUBMITS. Message lands
|
||||
# (subs==1) yet shipping reports exit 2. Must be exit 0.
|
||||
mk sub $'──────── \n'
|
||||
SINK="$TMP/sink.sub"
|
||||
out=$("$SEND" -L "$SOCKET" -t sub -m "echo x >>'$SINK'" 2>"$TMP/e.sub"); rc=$?; sleep 0.4
|
||||
if [ "$rc" = 0 ] && [ "$(subs "$SINK")" = 1 ]; then
|
||||
ok "glyphless: U+2500 box that submits => exit 0 (delivered, not 'UNDELIVERED')"
|
||||
else no "glyphless: U+2500 box that submits => exit 0" \
|
||||
"rc=$rc subs=$(subs "$SINK")(delivered=$([ "$(subs "$SINK")" -ge 1 ] && echo yes||echo no)) err=[$(cat "$TMP/e.sub")]"; fi
|
||||
|
||||
# --- D (duplicate arm): operator follows the rc=2 stderr and retries once. On the
|
||||
# glyphless box, shipping => two submissions (the reported duplicate). The
|
||||
# property: one logical send => exactly one submission. Same fix closes it.
|
||||
mk dup $'──────── \n'
|
||||
SINK="$TMP/sink.dup"
|
||||
tries=0
|
||||
for attempt in 1 2; do
|
||||
tries=$((tries+1))
|
||||
out=$("$SEND" -L "$SOCKET" -t dup -m "echo x >>'$SINK'" 2>/dev/null); rc=$?
|
||||
sleep 0.4
|
||||
[ "$rc" = 0 ] && break # operator stops retrying only when told delivered
|
||||
done
|
||||
if [ "$(subs "$SINK")" = 1 ]; then
|
||||
ok "duplicate: one logical send (rc-driven retry) => exactly one submission (tries=$tries)"
|
||||
else no "duplicate: one logical send => exactly one submission" "submissions=$(subs "$SINK") tries=$tries"; fi
|
||||
|
||||
# --- E (faithful hung managed TUI, NOT a cooked shell): raw/no-echo, paints nothing.
|
||||
# A cooked `sleep infinity` echoes the paste via the kernel line discipline and
|
||||
# false-passes a cursor-row fix that is correct on real seats (measured). So: raw.
|
||||
mk_rawstuck() { tmux new-session -d -s "$1" -x 120 -y 40 -c "$TMP" \
|
||||
"bash --noprofile --norc -c 'stty -echo -icanon min 1 time 0 2>/dev/null; exec sleep infinity'"; sleep 0.5; }
|
||||
mk_rawstuck estuck
|
||||
SINK="$TMP/sink.estuck"
|
||||
out=$("$SEND" -L "$SOCKET" -t estuck -r 1 -m "this stuck draft was never submitted" 2>/dev/null); rc=$?
|
||||
sleep 0.3
|
||||
if [ "$rc" != 0 ] && [ "$(subs "$SINK")" = 0 ]; then
|
||||
ok "raw/no-echo stuck TUI (not submitted) => non-zero (no false delivered)"
|
||||
else no "raw stuck TUI must NOT report delivered" "rc=$rc subs=$(subs "$SINK")"; fi
|
||||
|
||||
# --- F (busy/queued branch, your BUSY-not-runtime finding): glyphless pane rendering the
|
||||
# queued banner, never consuming. QUEUED_RE :113 fires before the glyph grep => rc=0.
|
||||
mk_busy() { tmux new-session -d -s "$1" -x 120 -y 40 -c "$TMP" \
|
||||
"bash --noprofile --norc -c 'printf \"Press up to edit queued messages\n\"; exec sleep infinity'"; sleep 0.5; }
|
||||
mk_busy ebusy
|
||||
SINK="$TMP/sink.ebusy"
|
||||
out=$("$SEND" -L "$SOCKET" -t ebusy -m "echo x >>'$SINK'" 2>/dev/null); rc=$?; sleep 0.3
|
||||
if [ "$rc" = 0 ]; then
|
||||
ok "busy/queued-banner glyphless => exit 0 (queued is delivery; runtime owns custody)"
|
||||
else no "busy/queued-banner must report delivered" "rc=$rc"; fi
|
||||
|
||||
# --- C (historical-bug guard): unresolvable target. No pane ever carried our draft
|
||||
# => must fail, never infer delivered from absence of a glyph/snippet.
|
||||
if out=$("$SEND" -L "$SOCKET" -t "nonexistent-$$" -m "echo x >>'$TMP/sink.wrong'" 2>/dev/null); then
|
||||
no "wrong-pane: unresolvable target must NOT report success" "expected non-zero, got 0"
|
||||
else ok "wrong-pane: unresolvable target => non-zero (no false delivered)"; fi
|
||||
|
||||
echo "---"; echo "pass=$pass fail=$fail"
|
||||
[ "$fail" = 0 ]
|
||||
@@ -4,10 +4,13 @@
|
||||
#
|
||||
# 1. DELIVERED — a REPL that renders a `❯ ` input box and submits on Enter
|
||||
# (text scrolls to history, box clears) => exit 0 "✓ delivered".
|
||||
# 2. UNCONFIRMED — a pane with NO locatable prompt glyph. This is the exact
|
||||
# historical FALSE POSITIVE: pre-patch it printed "✓ delivered"
|
||||
# exit 0; post-patch it MUST fail loud (exit 2, stderr
|
||||
# "could not confirm submission").
|
||||
# 2. DELIVERED — a pane with NO prompt glyph that DOES submit => exit 0. A pi
|
||||
# seat is this fixture (U+2500 rule, no glyph). Reshaped for
|
||||
# #1257; see the note at the fixture for why the old exit-2
|
||||
# assertion was wrong.
|
||||
# 2b. UNCONFIRMED— a glyphless pane that never submits (raw/no-echo hung TUI)
|
||||
# => must fail loud. This carries the historical
|
||||
# false-positive guard that fixture 2 used to be credited with.
|
||||
# 3. DRAFT — a `❯ `-prompt pane that never submits (message stays on the
|
||||
# input line) => exit 2, stderr "unsubmitted draft".
|
||||
set -uo pipefail
|
||||
@@ -37,19 +40,44 @@ else
|
||||
no "delivered: ❯-prompt REPL that submits => exit 0 ✓ delivered" "rc=$rc out=[$out] err=[$(cat "$TMP/e1")]"
|
||||
fi
|
||||
|
||||
# --- Fixture 2: NO prompt glyph (default bash PS1). THE regression: pre-patch this
|
||||
# was a silent false-positive "delivered"; post-patch it must be unconfirmed→exit 2.
|
||||
# --- Fixture 2: NO prompt glyph, and the pane DOES submit (interactive bash).
|
||||
# RESHAPED 2026-08-16 (#1257), deliberately. This fixture previously asserted
|
||||
# exit 2 here and was labelled "false-positive FIXED". That assertion was wrong,
|
||||
# and locking it in is what kept E7 alive: the pane submits, so "delivered" is
|
||||
# the truth, and a pi seat — whose input box is a bare U+2500 rule with no glyph
|
||||
# — IS this fixture. Reporting exit 2 for it told operators a delivered message
|
||||
# may be undelivered, and the retry that advice invites is the duplicate.
|
||||
#
|
||||
# The guard this fixture was reaching for is real and is NOT dropped: "never
|
||||
# infer delivered from absence" is now enforced positively by fixture 2b below
|
||||
# (glyphless AND not submitting => must fail) and by fixture 3 (locatable box
|
||||
# still carrying our tail => draft). Absence alone decides nothing either way.
|
||||
tmux -L "$SOCKET" new-session -d -s noglyph -c "$TMP" \
|
||||
'PS1="sh-noglyph$ " exec bash --noprofile --norc -i'
|
||||
sleep 0.3
|
||||
if out=$("$SEND" -L "$SOCKET" -t "=noglyph" -m "verdict fixture two must fail loud" 2>"$TMP/e2"); then
|
||||
no "unconfirmed: glyphless pane must NOT report success" "expected exit 2, got 0 (out=[$out])"
|
||||
out=$("$SEND" -L "$SOCKET" -t "=noglyph" -m "verdict fixture two must fail loud" 2>"$TMP/e2"); rc=$?
|
||||
if [ "$rc" -eq 0 ] && printf '%s' "$out" | grep -qF "✓ delivered"; then
|
||||
ok "delivered: glyphless pane that submits => exit 0 (runtime-agnostic, E7 FIXED)"
|
||||
else
|
||||
no "delivered: glyphless pane that submits => exit 0" "rc=$rc out=[$out] err=[$(cat "$TMP/e2")]"
|
||||
fi
|
||||
|
||||
# --- Fixture 2b: NO prompt glyph AND never submits — a hung managed TUI holding the
|
||||
# terminal in raw/no-echo, which is what a stuck agent seat actually is (measured
|
||||
# on live pi: stty -echo -icanon). Nothing is echoed, nothing is consumed, so
|
||||
# there is no positive evidence of submission and the tool MUST fail loud. This
|
||||
# is the historical false-positive guard, kept as a positive test.
|
||||
tmux -L "$SOCKET" new-session -d -s rawstuck -c "$TMP" \
|
||||
'bash --noprofile --norc -c "stty -echo -icanon min 1 time 0 2>/dev/null; exec sleep infinity"'
|
||||
sleep 0.3
|
||||
if out=$("$SEND" -L "$SOCKET" -t "=rawstuck" -r 1 -m "verdict fixture two-b never submitted" 2>"$TMP/e2b"); then
|
||||
no "unconfirmed: glyphless hung TUI must NOT report success" "expected non-zero, got 0 (out=[$out])"
|
||||
else
|
||||
rc=$?
|
||||
if [ "$rc" -eq 2 ] && grep -qF "could not confirm submission" "$TMP/e2"; then
|
||||
ok "unconfirmed: glyphless pane => exit 2 + 'could not confirm submission' (false-positive FIXED)"
|
||||
if [ "$rc" -ne 0 ] && grep -qF "could not confirm submission" "$TMP/e2b"; then
|
||||
ok "unconfirmed: glyphless hung TUI (raw/no-echo) => non-zero + 'could not confirm submission'"
|
||||
else
|
||||
no "unconfirmed: glyphless pane => exit 2 + stderr" "rc=$rc err=[$(cat "$TMP/e2")]"
|
||||
no "unconfirmed: glyphless hung TUI => non-zero + stderr" "rc=$rc err=[$(cat "$TMP/e2b")]"
|
||||
fi
|
||||
fi
|
||||
|
||||
|
||||
@@ -25,7 +25,7 @@
|
||||
"lint": "eslint src",
|
||||
"typecheck": "tsc --noEmit",
|
||||
"test": "vitest run --passWithNoTests && pnpm run test:framework-shell",
|
||||
"test:framework-shell": "bash framework/tools/quality/scripts/check-test-enumeration.sh && bash framework/tools/quality/scripts/test-check-test-enumeration.sh && python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/promotion_binding_unittest.py && python3 src/lease-broker/promotion_trigger_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/receipt_observer_client_unittest.py && python3 src/lease-broker/invariant_r_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-ci-queue-wait-tristate.sh && bash framework/tools/git/test-ci-queue-wait-github-checks.sh && bash framework/tools/git/test-pr-merge-queue-branch.sh && bash framework/tools/git/test-pr-merge-head-pin.sh && bash framework/tools/git/test-pr-merge-message-field.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/woodpecker/test-terminal-green-contract.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/_scripts/test-mosaic-init-rce.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh && bash framework/tools/glpi/test-list-http-status.sh && bash framework/tools/orchestrator/test-board-roll.sh && bash framework/tools/woodpecker/test-ci-wait-exit-matrix.sh && bash framework/tools/_scripts/test-fleet-transport-check.sh"
|
||||
"test:framework-shell": "bash framework/tools/quality/scripts/check-test-enumeration.sh && bash framework/tools/quality/scripts/test-check-test-enumeration.sh && python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/promotion_binding_unittest.py && python3 src/lease-broker/promotion_trigger_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/receipt_observer_client_unittest.py && python3 src/lease-broker/invariant_r_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-ci-queue-wait-tristate.sh && bash framework/tools/git/test-ci-queue-wait-github-checks.sh && bash framework/tools/git/test-pr-merge-queue-branch.sh && bash framework/tools/git/test-pr-merge-head-pin.sh && bash framework/tools/git/test-pr-merge-message-field.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/woodpecker/test-terminal-green-contract.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/_scripts/test-mosaic-init-rce.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh && bash framework/tools/fleet/test-start-agent-session.sh && bash framework/tools/glpi/test-list-http-status.sh && bash framework/tools/orchestrator/test-board-roll.sh && bash framework/tools/woodpecker/test-ci-wait-exit-matrix.sh && bash framework/tools/_scripts/test-fleet-transport-check.sh"
|
||||
},
|
||||
"dependencies": {
|
||||
"@mosaicstack/brain": "workspace:*",
|
||||
|
||||
Reference in New Issue
Block a user