spectre-capture
Capture durable project decisions, accepted corrections, verified reusable patterns or gotchas, disproved guidance, and changed…
Internal diagnosis engine for the Spectre Fix workflow. Use only when invoked by spectre-fix to investigate a reported bug and prepare its Execute handoff. Do NOT invoke directly for user requests.
$ npx -y skills add joenandez/spectre --skill spectre-fix-core --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/spectre-fix-coreContext preview
The summary Claude sees to decide when to auto-load this skill.
Internal diagnosis engine for the Spectre Fix workflow. Use only when invoked by spectre-fix to investigate a reported bug and prepare its Execute handoff. Do NOT invoke directly for user requests.
name: "spectre-fix-core" description: "Internal diagnosis engine for the Spectre Fix workflow. Use only when invoked by spectre-fix to investigate a reported bug and prepare its Execute handoff. Do NOT invoke directly for user requests." user-invocable: false
Own the diagnostic phase of the bug flow: reproduce the failure, ground the root cause and behavioral blast radius, then return a source-faithful Execute handoff. Execute owns all code changes.
Experience-contract rows map technical evidence to product behavior: `journey/surface | current experience | expected experience | technical path/consumer | disposition=intended-change|preserved-invariant|collateral-change|unresolved | evidence | verification`.
**DONE when:** root cause is grounded and its hypotheses were traced by dispatched analysts; the product and technical direct blast radius was independently explored after the candidate repair boundary was known; no row is `unresolved`; and the parent receives the diagnosis result.
1. **Diagnose.** Generate 5–7 hypotheses; reduce to 1–2 via evidence/data-flow/changes/paths; dispatch parallel `@spectre_analyst` traces; reproduce; synthesize one cause-level repair. 2. **Explore product + technical impact.** Once root cause and candidate repair boundary are grounded, dispatch ≥1 independent read-only `@spectre_analyst`; parallelize separable product journeys or technical boundaries. Trace shared callers, state, and data paths through to user/operator-observable outcomes. Return compact experience-contract rows that explicitly identify experiences that change, remain invariant, or are unresolved; synthesize and deduplicate the results. 3. **Honor phase.** `diagnose` returns `DIAGNOSIS_READY` before any code write. 3. **Contain scope.** Do not propose behavior beyond the reported repair boundary. An `unresolved` row or a new experience-contract row returns to the user for clarification.
Return phase status, hypotheses, diagnosis, experience contract with row-level evidence, proposed repair, limitations, and exact Execute source. Execute owns implementation and verification; the parent owns user-facing Next Steps.
SPECTRE is an Agentic Engineering workflow for Claude Code and Codex designed to help you do ONE THING more, faster, and with higher quality.
Capture durable project decisions, accepted corrections, verified reusable patterns or gotchas, disproved guidance, and changed…
Run final adversarial review through a pinned high-effort opposing runtime with one native fallback. Use at the final boundary to…
Generate a grounded draft pull request from the actual diff and open it via gh. Use when wrapping up a branch or writing a PR…
Assess and repair a project's Spectre knowledge, work history, tags, and retrieval quality through primary-led decisions,…
Execute tasks/plans with checks/reviews/proof. Use after planning/resume. Not for planning, unplanned fixes, or pruning.
Internal initializer for a missing Spectre feature root. Invoke only from a standalone workflow with no confirmed managed root.…