Notre avis
Exécute les commandes de build et de test configurées, capture les résultats et détermine si le code compile et les tests passent.
Points forts
- Vérification automatique et reproductible du build et des tests
- Capture des preuves dans un fichier de rapport structuré
- Gère les commandes manquantes ou les timeouts avec des statuts appropriés
Limites
- Nécessite une configuration préalable dans config.json
- Ne supporte que des commandes uniques (pas de chaînes complexes)
- Les sorties d'erreur peuvent être longues et difficiles à analyser
Utilisez cette porte après une modification du code pour valider que le projet reste compilable et les tests verts.
Ne l'utilisez pas pour des vérifications nécessitant une analyse humaine ou des validations non liées au build et aux tests.
Analyse de sécurité
PrudenceThe skill executes build and test commands defined in a config file using Bash, which could be abused if the configuration is compromised. However, it is intended as an internal gate and does not directly prompt for destructive or exfiltration actions.
- •Runs user-configurable shell commands via Bash; potential code injection if config.json is tampered with.
- •Uses Bash, a powerful tool that can execute arbitrary commands.
Exemples
Verify the build and tests to ensure the project compiles and all tests pass.Run the build gate to check if the code compiles, but skip the tests.Execute the test command defined in config to confirm the unit tests pass.name: gate-build description: >- Runs configured build and test commands. Captures output as evidence. Returns PASS if commands exit 0, FAIL otherwise. Internal gate — invoked by verify, not directly by users. allowed-tools:
- Read
- Bash
- Glob
- Write
Gate: Build
Goal
Confirm the codebase compiles and tests pass. This is the most basic gate — if the code doesn't build or tests don't pass, nothing else matters.
Inputs
.specwright/config.json--commands.buildandcommands.test.specwright/state/workflow.json-- current work unit for evidence path
Outputs
- Evidence file at
{currentWork.workDir}/evidence/build-report.md - Gate status update in workflow.json: PASS, FAIL, or ERROR
- Console output showing results inline (users see findings, not just badges)
Constraints
Execution (LOW freedom):
- Read build command from
config.jsoncommands.build. Run it. Capture output. - Read test command from
config.jsoncommands.test. Run it. Capture output. - If a command is null/missing, SKIP that check (not FAIL).
- If both commands are null, gate status is SKIP.
- Timeout: 5 minutes per command. If exceeded, status is ERROR.
Verdict (LOW freedom):
- Follow
protocols/gate-verdict.mdfor verdict rendering. - Build exit code 0 + test exit code 0 = PASS.
- Any non-zero exit code = FAIL.
- Show failing output inline so the user sees what broke.
Evidence (LOW freedom):
- Follow
protocols/evidence.mdfor file format and storage. - Write evidence file with: command run, exit code, stdout/stderr, timestamp.
- Update
workflow.jsongates section perprotocols/state.md.
Protocol References
protocols/gate-verdict.md-- default-FAIL, self-critique, visibilityprotocols/evidence.md-- evidence storage and freshnessprotocols/state.md-- gate status updates
Failure Modes
| Condition | Action | |-----------|--------| | Build command not configured | SKIP build check, continue to test check | | Test command not configured | SKIP test check | | Both commands null | Gate status = SKIP | | Command times out (>5min) | Gate status = ERROR with timeout message |
TDD Red-Green-Refactor
Testing
Skill qui guide Claude a travers le cycle TDD complet.
Audit d'Accessibilité Web
Testing
Réalise un audit d'accessibilité web complet selon les normes WCAG.
Générateur de Tests UAT
Testing
Génère des cas de test d'acceptation utilisateur structurés et complets.