rafaie/codex_home/smoke-test

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Run a work-item-scoped real entrypoint smoke test in provider-aware live or guaranteed offline mode, validate outputs and budgets, and emit standardized evidence. Use for runnable-path verification or smoke failure reproduction.

Install

npx vskill@latest install rafaie/codex_home/smoke-test

SKILL.md

1) Resolve the smoke command.

  • Read codex.toml and use [commands].smoke when present.
  • If absent, use uv run python scripts/smoke.py when scripts/smoke.py exists.
  • If neither exists, fail as config_error and report the missing command/script.

2) Scope smoke to the requested work item when provided.

  • Prefer --work-item <work-id> for scripts/smoke.py.
  • Treat --feature as a deprecated compatibility alias only.

3) Run smoke using the real entrypoint path.

  • Smoke must execute a subprocess command path (CLI/module/script), not only import-level calls.
  • Use tiny fixture inputs under tests/fixtures/smoke/ when available.
  • Pass the selected mode to the child as SMOKE_MODE; offline mode must not make provider network calls.
  • Select live mode only when credentials match a selected scenario provider, unless explicitly forced.

4) Enforce required artifact contract under artifacts/smoke/<run-id>/ (or equivalent configured directory):

  • summary.json
  • cases/<case_id>.json
  • stdout.txt
  • stderr.txt
  • timing.json

5) Validate outputs.

  • Always validate the configured JSON Schema, not merely JSON shape.
  • Always evaluate configured declarative invariants (required paths, equality, membership, and numeric ranges).
  • Golden comparison is optional per scenario and must support explicit --record updates.
  • Require child usage telemetry when token or cost ceilings are configured, and fail closed when a configured budget cannot be verified.

6) Summarize every run:

  • command run
  • work item ID (work_item) when provided
  • fixture(s)/scenario(s)
  • exit code
  • artifact locations
  • wall-clock time
  • budget telemetry (requests_used, time_budget_exceeded, max request/time settings)
  • source identity (Git commit/dirty diff hash, registry hash, fixture hashes, generated time)
  • failure bucket counts

7) Classify failures into one primary bucket per case/run summary:

  • config_error, fixture_error, auth, permission_error, network_error, timeout, rate_limit, quota_exceeded, provider_error, invalid_response, parse_error, schema_error, invariant_error, golden_mismatch, budget_exceeded, artifact_error, subprocess_error, internal_error

8) Use this exit-code mapping:

  • 0 pass
  • 10 config/fixture/auth/permission failure
  • 12 artifact failure
  • 20 subprocess failure
  • 22 network failure
  • 23 timeout
  • 24 rate limit
  • 25 quota exceeded
  • 26 provider error
  • 30 invalid/parse error
  • 40 schema validation failure
  • 41 invariants failure
  • 50 golden mismatch
  • 55 budget exceeded
  • 60 internal/unclassified

9) If multiple failures occur, prioritize exit code in this order:

  • artifact -> config/fixture/auth/permission -> quota -> rate_limit -> network -> timeout -> provider -> subprocess -> invalid/parse -> schema -> invariant -> golden -> budget -> internal

10) Recommend next action:

  • deterministic smoke/runtime/packaging issues -> debug-loop
  • fixture/golden updates only when behavior change is intentional and documented

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/rafaie/codex_home/smoke-test/badge)](https://verified-skill.com/skills/rafaie/codex_home/smoke-test)

Something wrong with this skill or its page? Tell us.