The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs several local shell operations to manage documentation branches and analyze changes.
Evidence: Uses git rev-parse, git checkout, git pull, git diff, and grep commands across several steps.
The DOCS_PATH variable and branch names are interpolated directly into shell strings, which is a common pattern for developer-oriented tools but requires the user to provide a trusted path.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes untrusted content from source code and documentation files.
Ingestion points: Reads source code files (e.g., src/pipecat/**/*.py), documentation files (.mdx), and git diff outputs in Step 5 and Step 7.
Boundary markers: The instructions do not define clear delimiters or include 'ignore embedded instructions' warnings for the data being analyzed.
Capability inventory: The agent has capabilities to execute local shell commands (git, grep), read files, and write/edit files on the local disk.
Sanitization: There is no evidence of sanitization or filtering of the content read from the source or doc files before the agent analyzes it to make editing decisions.

update-docs