The Agent Skills Directory

[COMMAND_EXECUTION]: The skill runs a local bash script located at .specify/scripts/bash/check-prerequisites.sh to parse project structure and available documentation.
[REMOTE_CODE_EXECUTION]: The skill processes an extension file (.specify/extensions.yml) and executes commands found under the hooks.before_critique and hooks.after_critique keys using the EXECUTE_COMMAND trigger. This allows for dynamic execution of arbitrary commands provided by the project being audited.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from project files.
Ingestion points: Reads project requirements from spec.md, implementation plans from plan.md, and configuration from .specify/extensions.yml.
Boundary markers: The skill lacks explicit markers or instructions to treat the content of these files as data rather than instructions.
Capability inventory: Includes shell script execution, arbitrary command execution (via hooks), and filesystem write operations for report generation.
Sanitization: No sanitization or validation is performed on the content of the ingested files before processing or command generation.

speckit-critique-run