← ClaudeAtlas

sb-dev-forward-testlisted

Run the behavioral forward tests for SpecBind's embedded skills against a fixture project, using subagents with no prior context. Use when a product-managed skill changed materially, before a release, or when a decision's effect on agent behavior needs checking.
Huruikagi/specbind · ★ 3 · Testing & QA · score 69
Install: claude install-skill Huruikagi/specbind
# Run the skill forward tests The [forward-test index](../../../docs/skill-forward-tests.md) routes to the scenario contracts, measurement dashboard, run archive, and findings worklist. The scenario documents are the contract; this skill is how to execute them without rediscovering the operational traps each time. The index also routes to the deliberately expensive end-to-end journey. Use that journey only for cross-lifecycle changes or release-candidate confidence; do not add it to the ordinary per-skill batch. This tests **product-managed skills** — the ones under `tools/specbind/assets/skills/` that get installed into consumer projects. It is not about this repository's own workflow. ## Before anything Record the commit under test: ```sh git rev-parse --short HEAD ``` A run measures one build. If you fix something mid-run, every scenario still in flight is measuring the old build, and its result is about that build only. Say so rather than folding it into the same table. ## One fixture per scenario ```sh sh tools/specbind/scripts/forward-test-fixture.sh /tmp/sb-<scenario> en ``` Never share a fixture between scenarios. Several depend on the starting state, and a leftover milestone silently changes what is being tested. The builder refuses to touch an existing directory, so delete and rebuild. The builder prints an `export PATH=` line for the CLI it ships inside the fixture. That line is not optional. The skills invoke `specbind` as a bare command because a real