The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/check.py script uses subprocess.run to execute local validation scripts such as validate_skill.py. These operations are limited to hardcoded paths for internal repository tools and are consistent with the skill's core function of auditing and reporting.\n- [PROMPT_INJECTION]: The skill processes data from multiple SKILL.md files, which constitutes a surface for indirect prompt injection.\n
Ingestion points: scripts/compat_matrix.py reads frontmatter from all markdown files located in the skills/ directory.\n
Boundary markers: The analysis does not use specific delimiters to separate untrusted content from internal logic, relying on the project's folder structure.\n
Capability inventory: Capabilities are limited to generating reports and executing local validation scripts via subprocess.run.\n
Sanitization: The skill uses yaml.safe_load to parse frontmatter, which prevents the execution of arbitrary code embedded in YAML tags.

skill-compat-matrix