aiw-failure-analysislisted

Reactive skill for when the agent has claimed a task complete and that claim is contradicted — by the user, by runtime behaviour, by manual verification, or by the agent itself. Use it whenever the user reports the behaviour is still broken, the fix didn't help, the system is doing the wrong thing, or runtime disagrees with the agent's tests or validation, with phrases like 'still broken' or 'didn't help'. Use it equally when the agent is about to supersede, replace, or redo its own recent attempt at the same defect, or to touch code its own recent change last touched for that defect, whoever noticed. Load this skill BEFORE proposing any fix. The trigger is a trust collapse: the core skills (aiw-ground-truth, aiw-testing, aiw-verification, aiw-validation, aiw-housekeeping) that should have caught the problem did not, so the agent cannot trust its oracle, tests, or verification until each is audited. It stops the agent jumping to the nearest plausible patch. It owns the structured pause after a contradicted 'd
philippe-ths/ai-coding-workflow · ★ 0 · AI & Automation · score 66

Install: claude install-skill philippe-ths/ai-coding-workflow

# Failure Analysis Read this file the moment a "done" claim is contradicted, by anyone, including yourself. Read it before any further code change, any retry request, and any "let me try X" suggestion. ## Why This Skill Exists When the agent declared the task complete, three things had to have signed off: the oracle was right, the tests passed, and verification concluded the change worked. If reality now disagrees, at least one of those three was wrong — and the agent has no immediate way to know which, because the agent was the actor in all three roles. This is a trust collapse. The agent's natural response is to jump to the nearest plausible fix. Resist it. If understanding was wrong enough for verification to miss the problem, the nearest plausible fix is likely also wrong, and a sequence of such fixes will produce confident-looking patches that move the broken behaviour around without resolving it. This skill is reactive. The core skills (aiw-ground-truth, aiw-testing, aiw-verification, aiw-validation, aiw-housekeeping) are proactive — they keep the loop honest. This skill runs when the loop failed anyway, and audits the core skills to find where the gap opened. ## Relationship to the Investigate Modality The Investigate modality (defined in aiw-ground-truth) covers tasks where producing understanding is the deliverable from the start — diagnosing a bug, narrowing down a cause, deciding what to build. Failure analysis covers tasks where the agent thought it had fin