The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a local shell script, scripts/new_prd.sh, designed to create and initialize PRD files. This script uses perl for string substitution, interpolating the PRD title directly into a command string, which presents a potential injection risk if the input contains special shell or regex characters.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its requirement to scan and read existing repository documentation.
Ingestion points: The workflow's 'Step 0' involves the agent searching for and potentially reading content from directories such as docs/, specs/, and adr/ to align with project conventions.
Boundary markers: No delimiters or safety instructions are provided to ensure the agent ignores malicious instructions that might be contained within existing project files.
Capability inventory: The skill is capable of executing local scripts, creating directories, and writing or copying files to the local file system.
Sanitization: Although the script generates a slug for the filename, the title string is not sanitized before being used in the perl in-place modification command.

prd-writer