self-assess-autopilotlisted

# self-assess-autopilot Run self-assess's full check -> plan -> fix -> validate value stream. ## Step 0: Settings gate ``` python3 ${CLAUDE_PLUGIN_ROOT}/scripts/self_assess_cli.py check-enabled --repo <repo_root> --skill self-assess-autopilot ``` ## Step 1: CHECK phase -- stage-map first, then everything else in parallel Rule `autopilot-stage-map-fresh-reuse`: **before** invoking `self-assess:self-assess-stage-map`, check whether its outputs can be reused instead of rebuilt: ``` python3 ${CLAUDE_PLUGIN_ROOT}/scripts/self_assess_cli.py stage-map-fresh-check --repo <repo_root> ``` - If `fresh` is `true`, **skip invoking `self-assess:self-assess-stage-map`** — its `stage_graph.json` and `file_stage_index.json` already exist and are not stale relative to the latest commit. Tell the user you're reusing the existing stage map rather than silently redoing it, and proceed straight to dispatching the other finding domains. - If `fresh` is `false`, invoke `self-assess:self-assess-stage-map` — either it has never run, a prior run left it incomplete, or the repo has changed since it last ran. This is also true the first time autopilot ever runs in a repo (nothing to reuse yet). Rule `autopilot-stage-map-first` (unchanged): whichever branch above applies, stage-map's outputs (fresh or reused) MUST be settled before any other finding domain starts — it writes `stage_graph.json` and `file_stage_index.json`, which `self-assess-arch-health` and `self-assess-transform-brief`