The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user and the local environment, which could potentially contain malicious instructions intended to influence the agent's logic during the PRD generation process.
Ingestion points: The skill captures a product idea directly from user input in Step 1 and reads various local files such as README.md, CLAUDE.md, and architecture documentation in Step 2.
Boundary markers: No explicit delimiters or instructions are provided to the agent to isolate the ingested data or to ignore any commands that might be embedded within the user idea or codebase files.
Capability inventory: The skill possesses the ability to read arbitrary project files, write new files to the ./prds/ directory, and execute the gh CLI to interact with GitHub.
Sanitization: The skill lacks explicit sanitization for the ingested content, although it instructions the agent to convert the user's idea into a kebab-case slug for use in titles and filenames.
[COMMAND_EXECUTION]: The skill utilizes the gh CLI to interact with a well-known service (GitHub).
The skill executes gh issue create --title "<slug>" --body-file ./prds/<slug>.md to push the generated PRD as a GitHub issue. This action is explicitly gated by a user confirmation prompt ([y/N]), providing a layer of manual oversight before the command is run.

shape