The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-controlled data, including documentation files and git logs, which creates a surface for indirect prompt injection where malicious instructions could be embedded in files analyzed by the agent. * Ingestion points: Project documentation (GDD.md, feature notes) and repository history (git diff, git log). * Boundary markers: Lacks explicit delimiters or instructions to ignore embedded content within ingested files. * Capability inventory: Reading and writing project files and executing a local Python script. * Sanitization: No sanitization or filtering logic for processed data was found.
[COMMAND_EXECUTION]: The skill executes a local Python script (scripts/check_docs.py) to validate documentation structure and link consistency as part of its verification workflow.

game-docs