fix(git-tools): harden gitea pr metadata wrappers

2026-05-22 10:17:39 -05:00
parent 755df9079e
commit a665c0ac4c
6 changed files with 457 additions and 52 deletions
--- a/docs/TASKS.md
+++ b/docs/TASKS.md
@@ -30,6 +30,7 @@ These are MVP-level checks that don't belong to any single workstream. Updated b
 | MVP-T04 | not-started | Sync `.mosaic/orchestrator/mission.json` MVP slot with this manifest (milestone enumeration, etc.)       | Coord state file; consider whether to repopulate via `mosaic coord` or accept hand-edit |
 | MVP-T05 | in-progress | Kick off W1 / FED-M1 — federated tier infrastructure                                                     | Session 16 (2026-04-19): FED-M1-01 in-progress on `feat/federation-m1-tier-config`      |
 | MVP-T06 | not-started | Declare additional workstreams (web dashboard, TUI/CLI parity, remote control, etc.) as scope solidifies | Track each new workstream by adding a row to the Workstream Rollup                      |
+| T-A292E96F | in-progress | Fix Mosaic Gitea PR metadata/login wrapper regression for U-Connect merge preflight | Kanban `t_a292e96f`; branch `fix/t-a292e96f-gitea-pr-metadata`; scratchpad `docs/scratchpads/t-a292e96f-gitea-pr-metadata.md` |

 ## Pointer to Active Workstream

--- a/docs/scratchpads/t-a292e96f-gitea-pr-metadata.md
+++ b/docs/scratchpads/t-a292e96f-gitea-pr-metadata.md
@@ -0,0 +1,53 @@
+# t_a292e96f — Gitea PR metadata wrapper fix
+
+## Objective
+
+Repair Mosaic git wrappers so Gitea PR metadata and merge preflight work for U-Connect PRs on `git.uscllc.com` without selecting the unrelated `git.mosaicstack.dev` tea login.
+
+## Findings
+
+- Reproduced the failure from `/src/uconnect-worktrees/t_39ce717c-authentik-smoke-gate` with the current `pr-metadata.sh`:
+  - PR #1905 returned JSON with `number=null`, `baseRefName=""`, `headRefName=""`.
+  - PR #1908 returned JSON with `number=null`, `baseRefName=""`, `headRefName=""`.
+- Root cause: the wrapper treated HTTP/API error payloads as PR payloads and normalized missing fields to empty strings.
+- The credential loader can return a non-working `git.uscllc.com` API token in this environment, while host-specific `~/.git-credentials` basic auth succeeds. The wrapper now falls back by host before normalization.
+- `tea login list` has only `git.mosaicstack.dev` configured here; `pr-merge.sh` previously forced `--login mosaicstack`, which is invalid for `git.uscllc.com` and caused `Login name mosaicstack does not exist`.
+
+## Changes
+
+- `packages/mosaic/framework/tools/git/detect-platform.sh`
+  - Added `get_gitea_basic_auth <host>` to retrieve host-specific HTTPS credentials from `~/.git-credentials` without printing secrets.
+- `packages/mosaic/framework/tools/git/pr-metadata.sh`
+  - Uses strict bash mode.
+  - Checks Gitea HTTP status and fails nonzero on API errors/non-JSON instead of emitting empty branch fields.
+  - Falls back from token auth to host-specific basic auth.
+  - Normalizes standard `head.ref`/`base.ref` and fallback branch fields.
+  - Requires non-empty `headRefName` and `baseRefName`.
+  - Preserves GitHub `gh pr view` behavior.
+- `packages/mosaic/framework/tools/git/pr-merge.sh`
+  - Reads metadata once for base-branch policy preflight.
+  - Selects a `tea` login only when its configured URL matches the repo host.
+  - Falls back to authenticated Gitea merge API when no matching `tea` login exists, avoiding the wrong `mosaicstack` login for USC repos.
+  - Keeps squash-only and main-only merge policy.
+- `packages/mosaic/framework/tools/git/test-pr-metadata-gitea.sh`
+  - Added fixture-based regression harness for standard Gitea fields, fallback branch fields, `refs/pull/<n>/head` plus `head.label` normalization, and API error payloads.
+
+## Documentation / changelog note
+
+This repository currently has no root `CHANGELOG.md`; the scratchpad and `docs/TASKS.md` carry the task-level change record for this wrapper fix.
+
+## Verification log
+
+- Red regression check: copied the new `test-pr-metadata-gitea.sh` harness next to `origin/main` wrapper scripts and ran it with `MOSAIC_TEST_WORK_DIR=$PWD/.mosaic-test-work/pr-metadata-gitea-red`; it failed as expected with `headRefName=''` and `baseRefName=''` on the fixture API-error path.
+- `bash -n packages/mosaic/framework/tools/git/{detect-platform.sh,pr-metadata.sh,pr-merge.sh,test-pr-metadata-gitea.sh}`: passed.
+- `shellcheck -x -P . -e SC1090 packages/mosaic/framework/tools/git/{detect-platform.sh,pr-metadata.sh,pr-merge.sh,test-pr-metadata-gitea.sh}`: passed.
+- `MOSAIC_TEST_WORK_DIR=$PWD/.mosaic-test-work/pr-metadata-gitea packages/mosaic/framework/tools/git/test-pr-metadata-gitea.sh`: passed; verifies standard Gitea fields, fallback branch fields, `refs/pull/<n>/head` label normalization, and nonzero API-error handling.
+- Installed wrapper parity: `/home/hermes/.config/mosaic/tools/git/{detect-platform.sh,pr-metadata.sh,pr-merge.sh}` byte-match the PR source copies after validation, so active U-Connect wrapper invocations use the same fix while source PR review runs.
+- Live sanitized U-Connect metadata from `/src/uconnect` with `MOSAIC_CREDENTIALS_FILE=/src/jarvis-brain/credentials.json`:
+  - PR #1905: `number=1905`, `baseRefName=main`, `headRefName=edith/t_39ce717c-authentik-smoke-gate`, `state=open`, `host=git.uscllc.com`.
+  - PR #1908: `number=1908`, `baseRefName=main`, `headRefName=fix/t_23fa9e1d-portal-health-backend`, `state=closed`, `host=git.uscllc.com`.
+- Merge preflight dry runs from installed wrappers:
+  - PR #1905: `Dry run: would merge PR #1905 on git.uscllc.com with authenticated Gitea API fallback (base=main, method=squash).`
+  - PR #1908: `Dry run: would merge PR #1908 on git.uscllc.com with authenticated Gitea API fallback (base=main, method=squash).`
+- PR: `https://git.mosaicstack.dev/mosaicstack/stack/pulls/518`, branch `fix/t-a292e96f-gitea-pr-metadata`.
+- CI: Recent PR/push pipelines failed before clone/test execution due Woodpecker/Kubernetes PVC API timeout: `dial tcp 10.43.0.1:443: i/o timeout`. No repository test step executed in CI; local targeted verification above remains clean.
--- a/packages/mosaic/framework/tools/git/detect-platform.sh
+++ b/packages/mosaic/framework/tools/git/detect-platform.sh
@@ -103,16 +103,28 @@ get_gitea_token() {
    if [[ -f "$cred_loader" ]]; then
        local token
        token=$(
+            # shellcheck source=/dev/null
            source "$cred_loader"
+            # Host-specific wrapper resolution must not inherit caller/global GITEA_*.
+            # load_credentials intentionally preserves existing env vars for interactive use,
+            # but metadata/merge wrappers need credentials matching the remote host.
+            unset GITEA_TOKEN GITEA_URL
            case "$host" in
                git.mosaicstack.dev) load_credentials gitea-mosaicstack 2>/dev/null ;;
                git.uscllc.com)      load_credentials gitea-usc 2>/dev/null ;;
                *)
+                    local matched=false
                    for svc in gitea-mosaicstack gitea-usc; do
-                        load_credentials "$svc" 2>/dev/null || continue
-                        [[ "${GITEA_URL:-}" == *"$host"* ]] && break
                        unset GITEA_TOKEN GITEA_URL
+                        load_credentials "$svc" 2>/dev/null || continue
+                        if [[ "${GITEA_URL:-}" == "https://$host" || "${GITEA_URL:-}" == "http://$host" || "${GITEA_URL:-}" == *"//$host" ]]; then
+                            matched=true
+                            break
+                        fi
                    done
+                    if [[ "$matched" != true ]]; then
+                        unset GITEA_TOKEN GITEA_URL
+                    fi
                    ;;
            esac
            echo "${GITEA_TOKEN:-}"
@@ -123,10 +135,12 @@ get_gitea_token() {
        fi
    fi

-    # 2. GITEA_TOKEN env var (may be set by caller)
+    # 2. GITEA_TOKEN env var (only when GITEA_URL, if present, matches the remote host)
    if [[ -n "${GITEA_TOKEN:-}" ]]; then
-        echo "$GITEA_TOKEN"
-        return 0
+        if [[ -z "${GITEA_URL:-}" || "${GITEA_URL:-}" == "https://$host" || "${GITEA_URL:-}" == "http://$host" || "${GITEA_URL:-}" == *"//$host" ]]; then
+            echo "$GITEA_TOKEN"
+            return 0
+        fi
    fi

    # 3. ~/.git-credentials file
@@ -143,6 +157,37 @@ get_gitea_token() {
    return 1
 }

+# Resolve HTTPS basic auth credentials for a Gitea host from ~/.git-credentials.
+# Prints "username:password" for direct curl -u consumption. Callers must not log it.
+get_gitea_basic_auth() {
+    local host="$1"
+    local creds="$HOME/.git-credentials"
+    if [[ ! -f "$creds" ]]; then
+        return 1
+    fi
+
+    python3 - "$host" "$creds" <<'PY'
+import sys
+from pathlib import Path
+from urllib.parse import unquote, urlparse
+
+host = sys.argv[1]
+creds = Path(sys.argv[2])
+
+for line in creds.read_text(encoding="utf-8").splitlines():
+    parsed = urlparse(line.strip())
+    if parsed.hostname != host:
+        continue
+    username = unquote(parsed.username or "")
+    password = unquote(parsed.password or "")
+    if username and password:
+        print(f"{username}:{password}")
+        raise SystemExit(0)
+
+raise SystemExit(1)
+PY
+}
+
 # If script is run directly (not sourced), output the platform
 if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
    detect_platform
--- a/packages/mosaic/framework/tools/git/pr-merge.sh
+++ b/packages/mosaic/framework/tools/git/pr-merge.sh
@@ -2,9 +2,10 @@
 # pr-merge.sh - Merge pull requests on Gitea or GitHub
 # Usage: pr-merge.sh -n PR_NUMBER [-m squash] [-d] [--skip-queue-guard]

-set -e
+set -euo pipefail

 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+# shellcheck source=packages/mosaic/framework/tools/git/detect-platform.sh
 source "$SCRIPT_DIR/detect-platform.sh"

 # Default values
@@ -12,6 +13,7 @@ PR_NUMBER=""
 MERGE_METHOD="squash"
 DELETE_BRANCH=false
 SKIP_QUEUE_GUARD=false
+DRY_RUN=false

 usage() {
    cat <<EOF
@@ -24,6 +26,7 @@ Options:
    -m, --method METHOD     Merge method: squash only (default: squash)
    -d, --delete-branch     Delete the head branch after merge
        --skip-queue-guard  Skip CI queue guard wait before merge
+        --dry-run           Run metadata/login preflight without merging
    -h, --help              Show this help message

 Examples:
@@ -54,6 +57,11 @@ while [[ $# -gt 0 ]]; do
            SKIP_QUEUE_GUARD=true
            shift
            ;;
+        --dry-run)
+            DRY_RUN=true
+            SKIP_QUEUE_GUARD=true
+            shift
+            ;;
        -h|--help)
            usage
            ;;
@@ -74,7 +82,8 @@ if [[ "$MERGE_METHOD" != "squash" ]]; then
    exit 1
 fi

-BASE_BRANCH="$("$SCRIPT_DIR/pr-metadata.sh" -n "$PR_NUMBER" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("baseRefName") or "").strip())')"
+PR_METADATA="$("$SCRIPT_DIR/pr-metadata.sh" -n "$PR_NUMBER")"
+BASE_BRANCH="$(printf '%s' "$PR_METADATA" | python3 -c 'import json, sys; print((json.load(sys.stdin).get("baseRefName") or "").strip())')"
 if [[ "$BASE_BRANCH" != "main" ]]; then
    echo "Error: Mosaic policy allows merges only for PRs targeting 'main' (found '$BASE_BRANCH')." >&2
    exit 1
@@ -92,21 +101,122 @@ PLATFORM=$(detect_platform)
 OWNER=$(get_repo_owner)
 REPO=$(get_repo_name)

+find_tea_login_for_host() {
+    local host="$1"
+    local logins_json
+    command -v tea >/dev/null 2>&1 || return 1
+    logins_json=$(tea login list --output json 2>/dev/null) || return 1
+    TEA_LOGINS_JSON="$logins_json" python3 - "$host" <<'PY'
+import json
+import os
+import sys
+
+host = sys.argv[1]
+try:
+    logins = json.loads(os.environ.get("TEA_LOGINS_JSON", "[]"))
+except Exception:
+    raise SystemExit(1)
+
+for login in logins if isinstance(logins, list) else []:
+    url = str(login.get("url") or login.get("URL") or "")
+    name = str(login.get("name") or login.get("Name") or "")
+    if url.rstrip("/").endswith(host) and name:
+        print(name)
+        raise SystemExit(0)
+
+raise SystemExit(1)
+PY
+}
+
+merge_gitea_with_api() {
+    local host="$1" api_url token basic_auth body_file raw_code payload
+    body_file=$(mktemp)
+    payload='{"Do":"squash"}'
+
+    token=$(get_gitea_token "$host" || true)
+    if [[ -n "$token" ]]; then
+        raw_code=$(curl -sS -w '%{http_code}' -o "$body_file" \
+            -X POST \
+            -H "Authorization: token $token" \
+            -H 'Content-Type: application/json' \
+            -d "$payload" \
+            "$api_url" || true)
+        if [[ "$raw_code" =~ ^2 ]]; then
+            rm -f "$body_file"
+            return 0
+        fi
+    fi
+
+    basic_auth=$(get_gitea_basic_auth "$host" || true)
+    if [[ -n "$basic_auth" ]]; then
+        raw_code=$(curl -sS -w '%{http_code}' -o "$body_file" \
+            -X POST \
+            -u "$basic_auth" \
+            -H 'Content-Type: application/json' \
+            -d "$payload" \
+            "$api_url" || true)
+        if [[ "$raw_code" =~ ^2 ]]; then
+            rm -f "$body_file"
+            return 0
+        fi
+    fi
+
+    python3 - "${raw_code:-000}" "$body_file" <<'PY' >&2
+import json
+import sys
+code, path = sys.argv[1], sys.argv[2]
+try:
+    data = json.load(open(path, encoding="utf-8"))
+    message = data.get("message") or data.get("error") or "unknown API error"
+except Exception:
+    message = open(path, encoding="utf-8", errors="replace").read()[:200] or "empty response"
+print(f"Error: Gitea API merge failed with HTTP {code}: {message}")
+PY
+    rm -f "$body_file"
+    return 1
+}
+
+if [[ "$DRY_RUN" == true ]]; then
+    if [[ "$PLATFORM" == "gitea" ]]; then
+        HOST=$(get_remote_host) || {
+            echo "Error: Cannot determine host from origin remote URL" >&2
+            exit 1
+        }
+        TEA_LOGIN="${GITEA_LOGIN:-$(find_tea_login_for_host "$HOST" || true)}"
+        if [[ -n "$TEA_LOGIN" ]]; then
+            echo "Dry run: would merge PR #$PR_NUMBER on $HOST with tea login '$TEA_LOGIN' (base=$BASE_BRANCH, method=squash)."
+        else
+            echo "Dry run: would merge PR #$PR_NUMBER on $HOST with authenticated Gitea API fallback (base=$BASE_BRANCH, method=squash)."
+        fi
+    else
+        echo "Dry run: would merge PR #$PR_NUMBER on $PLATFORM (base=$BASE_BRANCH, method=squash)."
+    fi
+    exit 0
+fi
+
 case "$PLATFORM" in
    github)
-        CMD="gh pr merge $PR_NUMBER --squash"
-        [[ "$DELETE_BRANCH" == true ]] && CMD="$CMD --delete-branch"
-        eval "$CMD"
+        GH_ARGS=(pr merge "$PR_NUMBER" --squash)
+        [[ "$DELETE_BRANCH" == true ]] && GH_ARGS+=(--delete-branch)
+        gh "${GH_ARGS[@]}"
        ;;
    gitea)
-        CMD="tea pr merge $PR_NUMBER --style squash --repo $OWNER/$REPO --login ${GITEA_LOGIN:-mosaicstack}"
+        HOST=$(get_remote_host) || {
+            echo "Error: Cannot determine host from origin remote URL" >&2
+            exit 1
+        }
+        TEA_LOGIN="${GITEA_LOGIN:-$(find_tea_login_for_host "$HOST" || true)}"
+        if [[ -n "$TEA_LOGIN" ]]; then
+            tea pr merge "$PR_NUMBER" --style squash --repo "$OWNER/$REPO" --login "$TEA_LOGIN"
+        else
+            echo "No tea login configured for $HOST; using authenticated Gitea API merge fallback." >&2
+            merge_gitea_with_api "$HOST" "https://${HOST}/api/v1/repos/${OWNER}/${REPO}/pulls/${PR_NUMBER}/merge"
+        fi

        # Delete branch after merge if requested
        if [[ "$DELETE_BRANCH" == true ]]; then
            echo "Note: Branch deletion after merge may need to be done separately with tea" >&2
        fi
-
-        eval "$CMD"
        ;;
    *)
        echo "Error: Could not detect git platform" >&2
--- a/packages/mosaic/framework/tools/git/pr-metadata.sh
+++ b/packages/mosaic/framework/tools/git/pr-metadata.sh
@@ -2,9 +2,10 @@
 # pr-metadata.sh - Get PR metadata as JSON on GitHub or Gitea
 # Usage: pr-metadata.sh -n <pr_number> [-o <output_file>]

-set -e
+set -euo pipefail

 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+# shellcheck source=packages/mosaic/framework/tools/git/detect-platform.sh
 source "$SCRIPT_DIR/detect-platform.sh"

 # Parse arguments
@@ -31,7 +32,7 @@ while [[ $# -gt 0 ]]; do
            exit 0
            ;;
        *)
-            echo "Unknown option: $1"
+            echo "Unknown option: $1" >&2
            exit 1
            ;;
    esac
@@ -42,56 +43,168 @@ if [[ -z "$PR_NUMBER" ]]; then
    exit 1
 fi

+write_metadata() {
+    local metadata="$1"
+    if [[ -n "$OUTPUT_FILE" ]]; then
+        printf '%s\n' "$metadata" > "$OUTPUT_FILE"
+    else
+        printf '%s\n' "$metadata"
+    fi
+}
+
+curl_gitea_pull() {
+    local api_url="$1"
+    local token basic_auth raw_code body_file http_code
+    body_file=$(mktemp)
+
+    token=$(get_gitea_token "$HOST" || true)
+    if [[ -n "$token" ]]; then
+        raw_code=$(curl -sS -w '%{http_code}' -o "$body_file" -H "Authorization: token $token" "$api_url" || true)
+        if [[ "$raw_code" =~ ^2 ]]; then
+            cat "$body_file"
+            rm -f "$body_file"
+            return 0
+        fi
+        http_code="$raw_code"
+    fi
+
+    basic_auth=$(get_gitea_basic_auth "$HOST" || true)
+    if [[ -n "$basic_auth" ]]; then
+        raw_code=$(curl -sS -w '%{http_code}' -o "$body_file" -u "$basic_auth" "$api_url" || true)
+        if [[ "$raw_code" =~ ^2 ]]; then
+            cat "$body_file"
+            rm -f "$body_file"
+            return 0
+        fi
+        http_code="$raw_code"
+    fi
+
+    if [[ -z "${http_code:-}" ]]; then
+        raw_code=$(curl -sS -w '%{http_code}' -o "$body_file" "$api_url" || true)
+        http_code="$raw_code"
+    fi
+
+    python3 - "$http_code" "$body_file" <<'PY' >&2
+import json
+import sys
+
+code, path = sys.argv[1], sys.argv[2]
+try:
+    data = json.load(open(path, encoding="utf-8"))
+    message = data.get("message") or data.get("error") or "unknown API error"
+except Exception:
+    message = open(path, encoding="utf-8", errors="replace").read()[:200] or "empty response"
+print(f"Error: Gitea pull request API request failed with HTTP {code}: {message}")
+PY
+    rm -f "$body_file"
+    return 1
+}
+
 detect_platform > /dev/null

 if [[ "$PLATFORM" == "github" ]]; then
    METADATA=$(gh pr view "$PR_NUMBER" --json number,title,body,state,author,headRefName,baseRefName,files,labels,assignees,milestone,createdAt,updatedAt,url,isDraft)
-
-    if [[ -n "$OUTPUT_FILE" ]]; then
-        echo "$METADATA" > "$OUTPUT_FILE"
-    else
-        echo "$METADATA"
-    fi
+    write_metadata "$METADATA"
 elif [[ "$PLATFORM" == "gitea" ]]; then
    OWNER=$(get_repo_owner)
    REPO=$(get_repo_name)
-    REMOTE_URL=$(git remote get-url origin 2>/dev/null)
-
-    # Extract host from remote URL
-    if [[ "$REMOTE_URL" == https://* ]]; then
-        HOST=$(echo "$REMOTE_URL" | sed -E 's|https://([^/]+)/.*|\1|')
-    elif [[ "$REMOTE_URL" == git@* ]]; then
-        HOST=$(echo "$REMOTE_URL" | sed -E 's|git@([^:]+):.*|\1|')
-    else
-        echo "Error: Cannot determine host from remote URL" >&2
+    HOST=$(get_remote_host) || {
+        echo "Error: Cannot determine host from origin remote URL" >&2
        exit 1
-    fi
+    }

    API_URL="https://${HOST}/api/v1/repos/${OWNER}/${REPO}/pulls/${PR_NUMBER}"
-
-    GITEA_API_TOKEN=$(get_gitea_token "$HOST" || true)
-
-    if [[ -n "$GITEA_API_TOKEN" ]]; then
-        RAW=$(curl -sS -H "Authorization: token $GITEA_API_TOKEN" "$API_URL")
+    if [[ -n "${MOSAIC_GITEA_PR_METADATA_RAW_FILE:-}" ]]; then
+        RAW=$(cat "$MOSAIC_GITEA_PR_METADATA_RAW_FILE")
    else
-        RAW=$(curl -sS "$API_URL")
+        RAW=$(curl_gitea_pull "$API_URL")
    fi

-    # Normalize Gitea response to match our expected schema
-    METADATA=$(echo "$RAW" | python3 -c "
-import json, sys
-data = json.load(sys.stdin)
+    # Normalize Gitea response to match GitHub's expected metadata schema.
+    METADATA=$(printf '%s' "$RAW" | python3 -c "
+import json
+import sys
+
+def first_non_empty(*values):
+    for value in values:
+        if value is None:
+            continue
+        if isinstance(value, str):
+            value = value.strip()
+        if value:
+            return value
+    return ''
+
+def nested(data, *keys):
+    current = data
+    for key in keys:
+        if not isinstance(current, dict):
+            return None
+        current = current.get(key)
+    return current
+
+try:
+    data = json.load(sys.stdin)
+except json.JSONDecodeError as exc:
+    print(f'Error: Gitea API returned non-JSON response: {exc}', file=sys.stderr)
+    sys.exit(1)
+
+if not isinstance(data, dict):
+    print('Error: Gitea API returned an unexpected non-object response', file=sys.stderr)
+    sys.exit(1)
+
+if data.get('message') and not data.get('number'):
+    print(f\"Error: Gitea API error: {data.get('message')}\", file=sys.stderr)
+    sys.exit(1)
+
+head_ref = first_non_empty(
+    nested(data, 'head', 'ref'),
+    nested(data, 'head', 'name'),
+    nested(data, 'head', 'branch'),
+    data.get('head_branch'),
+    data.get('head_ref'),
+    nested(data, 'head', 'label'),
+    data.get('head_label'),
+)
+if isinstance(head_ref, str) and head_ref.startswith('refs/pull/'):
+    head_ref = first_non_empty(
+        nested(data, 'head', 'label'),
+        data.get('head_label'),
+        nested(data, 'head', 'name'),
+        nested(data, 'head', 'branch'),
+        data.get('head_branch'),
+        data.get('head_ref'),
+        head_ref,
+    )
+base_ref = first_non_empty(
+    nested(data, 'base', 'ref'),
+    nested(data, 'base', 'name'),
+    nested(data, 'base', 'branch'),
+    data.get('base_branch'),
+    data.get('base_ref'),
+    data.get('base_label'),
+)
+
+if not head_ref or not base_ref:
+    available = ', '.join(sorted(data.keys()))
+    print(
+        'Error: Unable to resolve non-empty Gitea PR head/base refs '
+        f'(headRefName={head_ref!r}, baseRefName={base_ref!r}; keys={available})',
+        file=sys.stderr,
+    )
+    sys.exit(1)
+
 normalized = {
    'number': data.get('number'),
    'title': data.get('title'),
    'body': data.get('body', ''),
    'state': data.get('state'),
-    'author': data.get('user', {}).get('login', ''),
-    'headRefName': data.get('head', {}).get('ref', ''),
-    'baseRefName': data.get('base', {}).get('ref', ''),
-    'labels': [l.get('name', '') for l in data.get('labels', [])],
-    'assignees': [a.get('login', '') for a in data.get('assignees', [])],
-    'milestone': data.get('milestone', {}).get('title', '') if data.get('milestone') else '',
+    'author': nested(data, 'user', 'login') or '',
+    'headRefName': head_ref,
+    'baseRefName': base_ref,
+    'labels': [l.get('name', '') for l in data.get('labels', []) if isinstance(l, dict)],
+    'assignees': [a.get('login', '') for a in data.get('assignees', []) if isinstance(a, dict)],
+    'milestone': nested(data, 'milestone', 'title') or '',
    'createdAt': data.get('created_at', ''),
    'updatedAt': data.get('updated_at', ''),
    'url': data.get('html_url', ''),
@@ -102,11 +215,7 @@ normalized = {
 json.dump(normalized, sys.stdout, indent=2)
 ")

-    if [[ -n "$OUTPUT_FILE" ]]; then
-        echo "$METADATA" > "$OUTPUT_FILE"
-    else
-        echo "$METADATA"
-    fi
+    write_metadata "$METADATA"
 else
    echo "Error: Unknown platform" >&2
    exit 1
--- a/packages/mosaic/framework/tools/git/test-pr-metadata-gitea.sh
+++ b/packages/mosaic/framework/tools/git/test-pr-metadata-gitea.sh
@@ -0,0 +1,87 @@
+#!/usr/bin/env bash
+# Regression harness for Gitea PR metadata normalization.
+
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+WORK_DIR="${MOSAIC_TEST_WORK_DIR:-$PWD/.mosaic-test-work/pr-metadata-gitea}"
+REPO_DIR="$WORK_DIR/repo"
+FIXTURE_DIR="$WORK_DIR/fixtures"
+
+rm -rf "$WORK_DIR"
+mkdir -p "$REPO_DIR" "$FIXTURE_DIR"
+
+git -C "$REPO_DIR" init -q
+git -C "$REPO_DIR" remote add origin https://git.uscllc.com/USC/uconnect.git
+
+cat > "$FIXTURE_DIR/gitea-standard.json" <<'JSON'
+{
+  "number": 1905,
+  "title": "Smoke gate fix",
+  "state": "open",
+  "user": {"login": "edith"},
+  "head": {"ref": "edith/t_39ce717c-authentik-smoke-gate"},
+  "base": {"ref": "main"},
+  "labels": [{"name": "ci"}],
+  "assignees": [{"login": "edith"}],
+  "html_url": "https://git.uscllc.com/USC/uconnect/pulls/1905"
+}
+JSON
+
+cat > "$FIXTURE_DIR/gitea-fallback.json" <<'JSON'
+{
+  "number": 1908,
+  "title": "Fallback branch fields",
+  "state": "open",
+  "user": {"login": "edith"},
+  "head_branch": "fix/fallback-head",
+  "base_branch": "main",
+  "html_url": "https://git.uscllc.com/USC/uconnect/pulls/1908"
+}
+JSON
+
+cat > "$FIXTURE_DIR/gitea-refs-pull-label.json" <<'JSON'
+{
+  "number": 1908,
+  "title": "Closed merged PR with synthetic pull ref",
+  "state": "closed",
+  "user": {"login": "edith"},
+  "head": {"ref": "refs/pull/1908/head", "label": "fix/t_23fa9e1d-portal-health-backend"},
+  "base": {"ref": "main"},
+  "html_url": "https://git.uscllc.com/USC/uconnect/pulls/1908"
+}
+JSON
+
+cat > "$FIXTURE_DIR/gitea-error.json" <<'JSON'
+{"message": "user does not exist [uid: 0, name: ]", "url": "https://git.uscllc.com/api/swagger"}
+JSON
+
+run_case() {
+    local fixture="$1" expected_number="$2" expected_head="$3"
+    local output
+    output=$(cd "$REPO_DIR" && MOSAIC_GITEA_PR_METADATA_RAW_FILE="$fixture" "$SCRIPT_DIR/pr-metadata.sh" -n "$expected_number")
+    PR_METADATA_OUTPUT="$output" python3 - "$expected_number" "$expected_head" <<'PY'
+import json
+import os
+import sys
+
+data = json.loads(os.environ["PR_METADATA_OUTPUT"])
+expected_number = int(sys.argv[1])
+expected_head = sys.argv[2]
+assert data["number"] == expected_number, data
+assert data["baseRefName"] == "main", data
+assert data["headRefName"] == expected_head, data
+PY
+}
+
+run_case "$FIXTURE_DIR/gitea-standard.json" 1905 edith/t_39ce717c-authentik-smoke-gate
+run_case "$FIXTURE_DIR/gitea-fallback.json" 1908 fix/fallback-head
+run_case "$FIXTURE_DIR/gitea-refs-pull-label.json" 1908 fix/t_23fa9e1d-portal-health-backend
+
+if cd "$REPO_DIR" && MOSAIC_GITEA_PR_METADATA_RAW_FILE="$FIXTURE_DIR/gitea-error.json" "$SCRIPT_DIR/pr-metadata.sh" -n 1909 >/dev/null 2>"$WORK_DIR/error.log"; then
+    echo "Expected API error fixture to fail" >&2
+    exit 1
+fi
+grep -q "Gitea API error" "$WORK_DIR/error.log"
+
+echo "Gitea PR metadata regression harness passed"