The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the mkdir -p command to ensure the necessary directory structure exists before writing PDR files to the docs/prd/ path. This is documented in the Hard Rules and Execution Steps of SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input and existing local markdown files, creating a vulnerability to indirect instructions. 1. Ingestion points: User-provided product ideas and existing files at docs/prd/{module}/{area}/{slug}-pdr.md. 2. Boundary markers: The instructions do not specify any delimiters or safety prompts to prevent the agent from following instructions embedded within the ingested data. 3. Capability inventory: The skill has the ability to read files, create directories (mkdir), and write markdown files within the designated docs/prd/ workspace. 4. Sanitization: The skill applies kebab-case normalization to file slugs but does not sanitize the content of the PDR extracted from user input.

pdr-intake