Use this skill when a workflow is about to become a reusable agent artifact.
- Set
SKILL_ROOTto the absolute directory containing this installedSKILL.md. Do not assume the process working directory is the Bundle. - Set
TARGET_ROOTto the absolute working directory of the original workspace, and resolve candidate skill directories under that root. - Read
$SKILL_ROOT/references/contract.md, and validate the identifying fields of the portableSKILL.md. - Read
$SKILL_ROOT/references/decision-model.md, and distinguish repository context, reusable methods, external capabilities, lifecycle timing, deterministic logic, and isolated delegation. - Before execution, show the exact resolved argument vector. Run
python3 "$SKILL_ROOT/scripts/check_skill.py" "$TARGET_SKILL", whereTARGET_SKILLis the absolute directory of the candidate skill underTARGET_ROOT. - Inspect the JSON report. Fix all errors before discussing host-specific extensions.
- Compare the candidate artifact with
$SKILL_ROOT/assets/task-shapes.json, and return the minimal set of composable primitives.
Do not claim runtime extensions are part of the portable contract. Do not treat a valid skill as permission to run scripts or access tools.
Return the validation report, the selected primitives, and one sentence explaining each choice. Attach execution evidence: the resolved script path, target path, cwd, exact argv, and exit code. If the host cannot expose one of these observations, mark it as unverified; do not fabricate.