The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/validate_spec.py script utilizes subprocess.run to execute Git commands, including rev-parse, cat-file, and diff. These commands are used to verify that specification documents are properly committed to the repository and have not been modified post-approval, which is a legitimate functional requirement for maintaining approval integrity.
[COMMAND_EXECUTION]: The unit test script scripts/test_validate_spec.py employs subprocess.run to automate the initialization and configuration of a temporary Git repository for testing the validator's functionality.
[SAFE]: The skill processes untrusted content from natural language requests and repository files (Ingestion points: user input, docs/adr/ records, and specification files). While explicit prompt boundary markers are not used (Boundary markers: Absent), the skill mitigates risks associated with this attack surface by using a dedicated validation script that enforces a strict schema for frontmatter and validates document state against Git (Capability inventory: File writes to docs/specs/, shell execution of git; Sanitization: strict regex-based frontmatter validation and integrity checks).
[SAFE]: The test suite uses importlib.util to dynamically load the validate_spec.py module. This dynamic execution is safe as it is limited to a specific local file within the skill's own directory structure.

mission-spec