postmortem

Verify something about to be trusted by trying to refute it, in two separate judgments: construct the experiment that would prove it false (dispatched to the control-builder agent), then independently verify the attempt actually reached the subject before believing either outcome. Use when the user says 'adversarially verify this', 'build the control', 'try to refute this', 'prove this check can fail', 'did that green actually test anything', or when a new check, threshold, or 'this technique helps' belief is about to be trusted, or a result came back green where red was expected. Do NOT use for auditing a whole test suite (use test-audit) or for auditing prose claims (use claim-audit) — this is the single-claim primitive those audits dispatch to.

Census and live-fire audit of a repo's controls: everything check-shaped that fires outside the test suite (git hooks, Claude Code hooks, CLI validators, ambient reminders). Per control, four slots are re-derived from current code — fires-via, guarded-by, retirement-condition, disclosed-uncontrolled-edges — with 'nothing' in a slot as a reportable finding; controls nothing watches get deliberately violated on a scratch branch to confirm they fire. Use when the user says 'audit the controls', 'control audit', 'do our hooks actually fire', 'census the hooks', 'is anything watching this check', 'would the gate actually catch it'. Do NOT use for auditing a test suite (use test-audit) or prose claims (use claim-audit).

Evidence-grounded postmortem of finished work — a session, a feature, or a span of sessions. Recovers what went well, what did not, deviations from plan, and test escapes; every finding must cite a concrete artifact, and empty sections are valid output. Use when the user says 'postmortem', 'run a postmortem', 'retrospective', 'what went well', 'what would you do differently', or wants a look back at a completed plan, task, bugfix, feature, or run of sessions. Span mode mines git history, session logs, and changelogs. Do NOT use for auditing a test suite (use test-audit) or for drafting a session log.

Build a browsable HTML index of every postmortem in a repo — date, mode, scope, one-line conclusion, and the artifacts each one examined, plus a by-artifact view answering 'has anything been written about this file'. Reads frontmatter only; superseded postmortems are shown and marked, never hidden. Use when the user says 'browse postmortems', 'postmortem index', 'show me past postmortems', 'what have we written about X', 'list our retrospectives', or wants to find an earlier postmortem without knowing its filename. Do NOT use to run a postmortem (use postmortem) or to audit tests (use test-audit).

Audit an existing test suite for meaning and drift: recover each test's motivating claim, verify its oracle can actually fail (spot mutation — deliberately break the guarded behavior, confirm red), and map the reachability envelope (conditions the harness never exercises). Classifies tests as load-bearing, scar-tissue, decorative, or redundant and returns keep/rewrite/delete verdicts with evidence. Use when the user says 'audit the tests', 'test audit', 'are these tests testing the right thing', 'test drift', 'which tests are dead weight', 'do we trust this suite'. Do NOT use for writing new tests or for designing coverage on new features.