tunelisted

Tune the document at `$ARGUMENTS`. You are running the loop the human just asked for, so the runner may spend tokens. Report the estimate before the first call and the actual after. `${CLAUDE_PLUGIN_ROOT}` below is Claude Code's expansion for the plugin's install root. In a client that does not expand it, resolve it yourself: the plugin root is the directory three levels up from this SKILL.md — the one that contains `skills/`. The command blocks below use `<target-path>` as a slot, not a variable: replace it with the document path the human gave, keeping the surrounding double quotes so a path with spaces stays one argument. Never pass an unexpanded placeholder through to a shell. ## 1. Read it Read `$ARGUMENTS` in full, and the bundled doctrine at `${CLAUDE_PLUGIN_ROOT}/skills/skill-tuner/SKILL.md`. You need both: the probe tells you *what* is wrong, the doctrine tells you *what good looks like* when you fix it. ## 2. Probe it ``` python3 ${CLAUDE_PLUGIN_ROOT}/skills/skill-tuner/scripts/tune.py probe \ --target "<target-path>" --run-id tune-$(date +%s) --yes --budget-usd 3 --verify-trials 3 ``` Every finding it reports has already survived three independent skeptics in fresh contexts, plus a code-level check that the quoted text genuinely appears in the document. **Treat confirmed findings as real and act on them.** They are not suggestions to evaluate; the evaluating already happened. Read the run's `report.md` for each finding's quote and proposed fix. Refuted f