The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a shell command using string interpolation of user-provided input: python3 ../professional-brain/scripts/brain_query.py ./brain "<feature>". This pattern is vulnerable to command injection if the feature name contains shell metacharacters such as semicolons, pipes, or backticks.\n- [COMMAND_EXECUTION]: The instruction to save files using the user-provided feature name in the path entities/<feature>.md without validation presents a risk of directory traversal if the input contains path traversal sequences like ../.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes data from a sibling directory and user-provided inputs to generate documentation, creating a vulnerability surface where malicious content in those files could influence agent behavior.\n
Ingestion points: The skill reads from context.md, knowledge/strategy.md, and other feature files in the ../professional-brain/ directory.\n
Boundary markers: Absent. The instructions do not specify any delimiters or warnings to ignore embedded instructions within the ingested files.\n
Capability inventory: The skill has permissions to execute shell scripts and perform file system write operations.\n
Sanitization: Absent. There is no requirement to validate, escape, or filter content retrieved from the brain before using it in the PRD logic or logging it back to files.

prd-template