check

name	check
description	Run focused or complete validation for IronCrew. Use when the user asks to check, validate, lint, test, audit, or verify the repository state.
argument-hint	["focus|complete"]
disable-model-invocation	true
user-invocable	true
allowed-tools	Bash

Check

Choose the smallest focused tests while iterating, but finish every Rust task with the repository's complete all-target Rust gate.

Repository policy

Run the module policy for production Rust changes and the repository validators for every changed non-Rust policy surface:

• python3 -B scripts/check_module_size.py
• python3 -B -m unittest discover -s scripts/tests -p 'test_*.py'
• bun run scripts/validate_skills.ts
• bun run scripts/issues_registry.ts check
• bun test scripts/tests/*.test.ts
• actionlint .github/workflows/*.yml after workflow changes, when available
• bun run scripts/check_worktree.ts
• git status --short, with explicit review of all tracked and untracked source, policy, and documentation changes

Required Rust gate

Run in order and stop on the first failure:

1. cargo fmt --all -- --check
2. cargo clippy --all-targets -- -D warnings
3. cargo test --all-targets
4. cargo test --doc after public Rust documentation changes

Run cargo audit --deny warnings after dependency or security-sensitive changes.

Add ./scripts/check-lua-examples.sh for Lua/docs/example changes. Add the crew-effectiveness or replica-soak Python unit/contract checks for evaluation changes. For storage, idempotency, leases, HITL, journals, or replica behavior, run the serial PostgreSQL 15 integration targets with a disposable database and IRONCREW_TEST_PG_URL; a skipped test is not evidence.

Never use shared/production storage or print .env values. Report commands as pass, fail, skipped, or not applicable and distinguish local, PostgreSQL, process-level, soak, live-provider, Railway, and OpenShift evidence. Do not modify failures unless the user also asked for fixes.