agentleFS
Ar9avAr9av/game-exa/skills/parallel-qa-orchestra/SKILL.md

parallel-qa-orchestra

Skill4 starsChanged 7 months agoMITView on GitHub

Replaces the serial playtester → gap-checker chain with 4 concurrent QA workers: static analysis (no browser), golden-path scenario runner, dynamic fuzzer, and physics debug validator. Merges all results into a single failure list for the refiner. Cuts QA time from ~2 min to ~30s.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All 4 run concurrently. Total wall-clock time ≈ max(individual times) ≈ 30–45s.

    line 23

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Parallel QA Orchestra — Concurrent Validation

Orchestrates 4 independent QA checks that run simultaneously, each on its own dev-server port. The refiner gets a single merged failure report with every issue in one pass, cutting the number of refine-iterate cycles needed.

When to use

Replace the sequential playtester → gap-checker calls in the main pipeline with a single call to this skill. Use after codesmith and the manifest-auditor + mobile-compat-checker linting passes.

The 4 workers

WorkerBrowser?What it catches
staticNoBFS reachability, border holes, jump-arc gaps, spawn collision
golden-pathYes (port A)Boot, walk, jump, attack scenarios; win-condition reachability
fuzzerYes (port B)Stuck state, spawn-trap, no-progress, out-of-bounds, NaN state
physics-debugYes (port C)Hitbox misalignment, body offset bugs

All 4 run concurrently. Total wall-clock time ≈ max(individual times) ≈ 30–45s.

Output contract

{
  "ts": "<ISO>",
  "ok": false,
  "failCount": 3,
  "workers": {
    "static":       { "ok": true,  "errors": 0, "warnings": 1, "issues": [...] },
    "golden-path":  { "ok": false, "passed": false, "failures": [...] },
    "fuzzer":       { "ok": true,  "issues": [] },
    "physics-debug": { "ok": true, "misalignedCount": 1, "bodies": [...] }
  },
  "merged": [
    { "kind": "border-hole",         "source": "static",  "severity": "error",   "message": "..." },
    { "kind": "screenshot-diff",     "source": "golden",  "severity": "error",   "message": "..." },
    { "kind": "hitbox-misaligned",   "source": "physics", "severity": "warning", "message": "..." }
  ]
}

Written to qa/orchestra-report.json. Also appended to state.qa[] (keeping last 5).

Exit code 0 if ok, else 5.

Port allocation

Each browser worker gets its own vite dev-server port. The script finds 3 free ports automatically using Node's net module before launching workers.

Integration with the pipeline

Replace steps 7–8 in the gameforge orchestrator:

// Before:
playtester → refiner (loop) → playtester (loop)

// After:
manifest-auditor → mobile-compat-checker → parallel-qa-orchestra → refiner (loop) → parallel-qa-orchestra (loop)

Max 3 refiner iterations, same as before.

Process

  1. Run scripts/orchestrate_qa.mjs <project-dir>
  2. Script allocates 3 free ports.
  3. Spawns 4 workers concurrently via Promise.all + child_process.spawn.
  4. Collects stdout JSON from each worker as it finishes.
  5. Merges into orchestra-report.json.
  6. Exits with combined pass/fail.

Scripts

  • scripts/orchestrate_qa.mjs <project-dir> [--no-physics] — run all 4 workers. --no-physics skips the physics-debug worker (faster, loses hitbox checks).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.