The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill includes a utility script that dynamically creates new files and assigns them executable permissions.
Evidence: example_script.chmod(0o755) in scripts/init_skill.py is used to make generated template scripts runnable.
[INDIRECT_PROMPT_INJECTION]: The validation logic processes external Markdown and YAML files, creating a surface for instructions embedded in data to potentially influence the agent's behavior during the validation workflow.
Ingestion points: scripts/quick_validate.py reads SKILL.md and referenced files from a specified directory using Path.read_text().
Boundary markers: The script relies on standard YAML frontmatter delimiters (---) and Markdown heading patterns.
Capability inventory: The skill has the capability to create directories and files (init_skill.py), read file contents (quick_validate.py), and package files into archives (package_skill.py).
Sanitization: The script correctly uses yaml.safe_load() to prevent unsafe deserialization of YAML data.
[COMMAND_EXECUTION]: The workflow instructions direct the agent to execute several Python-based validation and testing scripts.
Evidence: SKILL.md contains instructions to run scripts/validate_skill_engine.py, scripts/routing_smoke_test.py, and scripts/quick_validate.py.

skill-writing