The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze source code, interaction states, and tests from an external repository provided by the user to generate documentation (Phase 0, Phase 2, Phase 3). This creates a surface for indirect prompt injection where malicious content within the source repo could attempt to influence the agent's behavior or output.
Ingestion points: The agent is instructed to read interaction state (state machines, reducers), domain objects, UI code, and behavioral tests from the user's source repository (README.md, goal-template.md).
Boundary markers: The instructions do not specify any delimiters or safety markers to isolate the ingested code from the agent's operational instructions.
Capability inventory: The agent performs file system writes to create the documentation repo, executes a local Python script (check-links.py), and runs git commands.
Sanitization: There is no evidence of sanitization or filtering of the source code before it is processed by the agent.
[COMMAND_EXECUTION]: The skill requires the execution of standard system commands to function, specifically git for repository management and python3 to run the included check-links.py script for consistency checks. These commands are localized to the documentation repository and the skill's own reference directory.

pe-product-description