The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains internal test scripts (tools/tests/smoke_test_skeleton.py and tools/tests/smoke_test_trigger_score.py) that use subprocess.run to execute other scripts within the same skill package. These executions are limited to testing the skill's own validator tools using the local Python interpreter.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves auditing and modifying other agent skill files, which creates an inherent surface for indirect prompt injection if the source data contains malicious instructions.
Ingestion points: Target skill directories and SKILL.md files read by the tools/skeleton.py utility.
Boundary markers: The skill enforces strict structure based on the assets/skill-template.md and uses specific Markdown heading patterns to delimit processing areas.
Capability inventory: File system read/write access for skill management and local script execution via subprocess for automated testing.
Sanitization: Implements identifier validation using regex (e.g., KEBAB_RE) and checks for key-name drift in frontmatter to prevent unauthorized data structures.

yzr-skill-management