tq-forge-testlisted

# /tq-forge-test — re-validate without rewriting ## When to use You manually edited a sandbox or production artifact and want to confirm it still passes. Or `forge-queue.json` has a slug in `needs_manual_review` and you want to know if today's edit cleared the bar. This skill never rewrites the artifact — only the score. ## Procedure ```bash export TQ_FORGE_HOME="${TQ_FORGE_HOME:-$HOME/.tq-forge}" S="${CLAUDE_PLUGIN_ROOT:-${TQ_FORGE_HOME:-$HOME/.tq-forge}/install}/scripts" source "$S/common.sh" && tq_ensure_home ``` 1. **Resolve the target.** Accept a slug or a full path. With a slug, auto-detect across sandbox + production: ```bash bash "$S/quality-score.sh" --auto-detect "<slug>" ``` If missing everywhere, list the sandbox + production dirs and abort: ```bash ls "$SANDBOX_SKILLS" "$SANDBOX_AGENTS" "$CLAUDE_SKILLS_DIR" 2>/dev/null ``` 2. **Run the structural dry-test** (resolve the path first via `tq_resolve`): ```bash P="$(tq_resolve "<slug>")"; bash "$S/dry-test.sh" --json "$P" ``` Capture verdict (`pass`/`warn`/`fail`), checks_passed, and `issues`. 3. **Run quality-score.** ```bash bash "$S/quality-score.sh" --json "$P" ``` Capture avg and per-dimension breakdown. 4. **Print a combined report.** Two-line header, then the per-dimension bars, then any dry-test `issues` as bullets: ``` <emoji> <slug> (<kind>) — quality X/10 · dry-test Y/10 (<verdict>) ``` 5. **Update skill-log.json** with the new score (mut