The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute docker compose commands that run a skill-tester container. This container executes dynamically generated Python test files using pytest.
[DYNAMIC_EXECUTION]: The skill generates executable Python scripts (e.g., test/tests/test_{snake_name}.py) by interpolating user-provided scenario descriptions, judge questions, and prompt strings into a predefined template. These scripts are subsequently executed in a shell environment via Docker.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes external SKILL.md files from the dstoic/skills/ directory to generate "golden fixtures" and test scaffolds. There are no explicit sanitization or boundary markers defined to prevent malicious instructions within these external files from influencing the generated code or agent behavior during the testing process.
Ingestion points: Reads dstoic/skills/{skill_name}/SKILL.md to generate simplified versions and test files.
Boundary markers: None identified for external content ingestion.
Capability inventory: Uses Bash, Write, and Edit tools; executes Docker commands and generates Python scripts.
Sanitization: No evidence of input validation or escaping for the content read from external skills.

test-skill