feat(recipe): add progress reporting banners and progress file lifecycle

[rysweet]

Summary

• Emit [recipe-runner] Preparing recipe '<name>'... and [recipe-runner] Launching recipe '<name>' (pid N)... banners to stderr so parent sessions can observe subprocess activity
• Add _cleanup_progress_file() to remove progress files on recipe completion (best-effort, catches OSError)
• Update existing progress file test to capture writes before cleanup removes the file

Redo of closed PR #3473.

Files changed (commit-only)

File	Change
src/amplihack/recipes/rust_runner.py	Preparing/Launching banners, _cleanup_progress_file(), cleanup in finally block
tests/recipes/test_rust_runner.py	New - 12 tests: banners (3), progress file lifecycle (6), get_recipe_progress integration (3)
tests/recipes/test_rust_runner_execution.py	Updated test_progress_mode_writes_progress_file to intercept writes before cleanup

No changes to models.py, parser.py, pyproject.toml, uv.lock, or any YAML recipe files.

What already existed (no changes needed)

• _write_progress_file() and _stream_process_output_with_progress() in rust_runner_execution.py
• get_recipe_progress() in dev_intent_router.py

Test evidence

tests/recipes/test_rust_runner.py ............                           [ 34%]
tests/recipes/test_rust_runner_execution.py .......................      [100%]
============================== 35 passed in 0.14s ==============================

All pre-commit hooks pass (ruff, pyright, ruff-format, secrets detection, print statement check).

Test plan

• uv run python -m pytest tests/recipes/test_rust_runner.py -x -q -- 12 pass
• uv run python -m pytest tests/recipes/test_rust_runner_execution.py -x -q -- 23 pass
• Pre-commit hooks green
• git diff-tree --name-only -r HEAD shows only 3 files
• CI green
• qa-team tests
• quality-audit 3 cycles

🤖 Generated with Claude Code

[github-actions]

Repo Guardian - Passed

All changed files are durable project artifacts — no ephemeral content detected.

File	Type	Assessment
src/amplihack/recipes/rust_runner.py	Source code	✅ Core module — durable
tests/recipes/test_rust_runner.py	Test file	✅ New test suite — durable
tests/recipes/test_rust_runner_execution.py	Test file	✅ Updated tests — durable

No action required.

Generated by Repo Guardian for issue #3487 · ◷

[rysweet]

PR #3487 Merge Validation Report

1. Scope Check: PASS

Exactly 3 files changed (verified via git diff origin/main...origin/fix/recipe-progress-reporting --name-only):

• src/amplihack/recipes/rust_runner.py (+21 lines)
• tests/recipes/test_rust_runner.py (+337 lines, new file)
• tests/recipes/test_rust_runner_execution.py (+31/-10 lines)

No changes to models.py, parser.py, pyproject.toml, or uv.lock. Confirmed clean.

2. QA-Team Outside-In Tests: PASS

Wrote and executed 2 gadugi YAML scenarios with a Python test harness (test_recipe_progress_reporting.py, 8 tests):

Scenario 1 — Progress banners to stderr (recipe-progress-banner-stderr.yaml):

• [recipe-runner] Preparing recipe '<name>'... emitted to stderr
• [recipe-runner] Launching recipe '<name>' (pid <pid>)... emitted to stderr
• Preparing appears before Launching (ordering verified)
• Banners go to stderr, NOT stdout (verified explicitly)

Scenario 2 — Progress file cleanup (recipe-progress-file-cleanup.yaml):

• _write_progress_file creates valid JSON in tempdir
• _cleanup_progress_file removes the file (no orphans)
• Cleanup is no-op when file already absent (no exception)
• Cleanup runs on success (finally block)
• Cleanup runs on failure (finally block)

tests/gadugi/test_recipe_progress_reporting.py ........    [8 passed in 0.10s]

3. Quality Audit — 3 Cycles: PASS

Cycle 1 (SEEK/VALIDATE):

• All 35 PR tests pass (test_rust_runner.py 12 + test_rust_runner_execution.py 23)
• Ruff lint: All checks passed
• except OSError: pass in _cleanup_progress_file is acceptable — best-effort cleanup in a finally block, not business logic error swallowing
• Minor DRY concern: name sanitization logic duplicated between _cleanup_progress_file and _progress_file_path, but acceptable to avoid cross-module coupling for a simple cleanup function

Cycle 2 (SEEK/VALIDATE — edge cases & security):

• Path traversal: re.sub(r"[^a-zA-Z0-9_]", "_", stem)[:64] sanitization prevents all injection
• Empty names, null bytes, long names: all handled safely
• No security issues found

Cycle 3 (SEEK/VALIDATE — integration):

• All recipe tests pass (35/35)
• Pre-existing failure in test_default_workflow_step19d_step21.py is unrelated (tests recipe YAML not touched by this PR)
• No regressions in any PR-related tests
• No FIX needed in any cycle

4. CI Status: PASS (with documented pre-existing exception)

Check	Status
Validate Code	PASS
GitGuardian Security Checks	PASS
Root Directory Hygiene Checks	PASS
Check Version Bump	PASS
Check skill/agent drift	PASS
Bot Detection	PASS (SKIPPED — not a bot)
Dependabot Auto-Merge	PASS
Repo Guardian	PASS
Claude Code Plugin Test	PASS
Invisible Character Scan	SKIPPED
Atlas PR Impact Check	FAIL (pre-existing)

Atlas failure explanation: The ast-lsp-bindings layer is stale because rust_runner.py was modified. However, the atlas was already stale on main after PR #3465 (fix(recipes): report workflow step progress) was merged without an atlas rebuild. This is NOT caused by this PR's changes — it's a pre-existing staleness condition. The atlas check would fail on any PR touching rust_runner.py until the atlas is rebuilt on main.

5. Verdict

READY TO MERGE

All merge criteria satisfied:

• Scope: exactly 3 files, no unwanted changes
• QA outside-in: 8 gadugi scenario tests pass
• Quality audit: 3 cycles, 0 issues requiring fix
• CI: all checks green except pre-existing Atlas staleness
• Test evidence: 35 PR tests + 8 gadugi tests = 43 total, all passing

[github-actions]

PR Triage Report

Risk: Low | Priority: Medium

Assessment

• Type: Feature — progress reporting enhancement for recipe runner
• Scope: 3 files — rust_runner.py (banners + cleanup), new test_rust_runner.py (12 tests), updated test_rust_runner_execution.py
• CI Status: Passing — Validate Code ✅, GitGuardian ✅ (Atlas PR Impact Check failure is non-blocking/expected)
• Test Coverage: 35 tests pass (12 new + 23 existing)

Findings

• Clean, focused scope — redo of feat(progress): recipe runner progress reporting — banners, progress files, and YAML markers #3473 with the same intent
• Progress reporting banners emitted to stderr so parent sessions can observe subprocess activity
• Progress file cleanup handled in finally block (best-effort, catches OSError)
• No changes to models.py, parser.py, pyproject.toml, uv.lock, or YAML recipe files

Recommendation

✅ Ready for review and merge. Low risk, well-tested, clear scope. Unblocks parent session observability.

Generated by PR Triage Agent · ◷