Files

T

jason.woltjeandClaude Opus 5.5 f539466fcb feat(queue): Piece D, reviews as issue comments, raw per-seat token helper (row 12, #1508 )

queue move ID in-review posts the review request as a Gitea comment and
review record reads verdicts back, so reviews stop being files in
docs/plans/reviews/. On a comment round, in-review to waiting-on-jason
now needs every listed reviewer's approval for the current round, the
same as in-review to done (Filbert r1 C1). scripts/gitea-api.sh reads
the raw per-seat token files (lead decisions 37 to 39): config built and
checked before curl starts, export attribute cleared, fixed base URL.
test-queue.sh skips its live checks outside the canonical root.

Darkwing authored. Filbert approved D r2 (cf1d3fd0) after r1 (a2dc2302)
and corrected the plan (293747cd). Rocko reviewed the helper (e896192f,
2096b0a3), and Sage's lead check passed under decision 38. Manifest
b402fb38, 19 files.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

2026-09-27 10:07:29 -05:00

work

feat(queue): Piece D, reviews as issue comments, raw per-seat token helper (row 12, #1508 )

2026-09-27 10:07:29 -05:00

CONTEXT.md

docs(agents): seats read the queue, not CURRENT.md (row 10, #1508 )

2026-09-26 20:23:20 -05:00

launch.sh

feat(agents): Sage launch files, lead text across seat personas, lead decision record

2026-09-26 15:11:10 -05:00

README.md

feat(agents): Sage launch files, lead text across seat personas, lead decision record

2026-09-26 15:11:10 -05:00

SOUL.md

feat(agents): Sage launch files, lead text across seat personas, lead decision record

2026-09-26 15:11:10 -05:00

README.md

Rocko development TUI

Run from any terminal:

/mnt/storage/src/mosaic-stack/agents/rocko/launch.sh

Rocko handles general Mosaic Stack development assignments under project lead Sage. The launcher opens Claude Code in the repository with --model sonnet and the display name Rocko. Sonnet is the selected CLI model alias; Rocko's project persona is supplied by SOUL.md, not a separate custom Claude agent.

Use --check to validate the executable and context inputs without a model call. This does not verify authentication or account access to Sonnet. Claude Code uses its normal authentication, settings, permissions, and discovered resources; this script does not copy credentials or bypass permissions.

Normal launches resume the exact Claude UUID recorded locally in .pi/state/rocko/session-id. First launch creates a UUID; --fresh creates another while retaining previous conversations in Claude's native storage. Each private launch receipt under .pi/state/rocko/launches/ retains its UUID and context snapshot. The launcher lock refuses concurrent Rocko launches through this script; it does not lock sessions opened by other means.

If Claude exits before saving its first conversation, or its history is removed, the saved UUID may not be resumable. Inspect the failure and use --fresh explicitly to start again. The launcher does not fall back to the latest unrelated project conversation. Old UUIDs in launch receipts can be opened manually with claude --resume UUID from the project directory.

The prompt snapshot includes CONSTITUTION, STANDARDS, Rocko's SOUL, the live Mosaic user profile, root AGENTS.md, and Rocko's CONTEXT.md. --soul FILE, --constitution FILE, and --user FILE override individual inputs; relative paths resolve from the repository root. Missing or empty inputs refuse launch. The snapshot is appended to Claude's system prompt on each launch.

This is a host development session, not a managed fleet registration. It loads no Pi goal extension. The user or team lead supplies the assignment.