speckit-fix-findings-run

What problem does it solve?

It resolves discrepancies between your specification (spec.md/plan.md) and the current implementation by repeatedly analyzing, fixing, and re-analyzing until no actionable findings remain.

Core Features & Use Cases

• Analyze–Fix–Reanalyze loop: Runs an automated cycle that starts from existing analysis (analysis.md) or generates it, then applies code fixes for actionable findings, and re-runs analysis to confirm resolution.
• Spec-aligned guardrails: Applies fixes only when they clearly align with spec.md and plan.md, skips ambiguous items, and never modifies spec artifacts (spec.md, plan.md, tasks.md).
• Findings auditing and traceability: Produces and updates a findings.fixed.md log with per-iteration resolution/deferment details and a final status (CLEAN, DEFERRED_REMAINING, or MAX_ITERATIONS_REACHED).
• Use Case: When your CI fails or your spec-kit analysis reports multiple critical and warning issues, use this skill to reduce the failures quickly and leave only deferred, human-decision items.

Quick Start

Run speckit-fix-findings-run from the repo root after ensuring the specified feature directory contains spec.md, plan.md, and tasks.md so the skill can locate analysis.md and apply iterative fixes.