The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill relies on processing external and potentially untrusted data sources, including product requirements documents (PRDs), roadmap notes, and issue discussions. This creates a surface for indirect prompt injection where malicious instructions embedded in these sources could attempt to influence the agent's behavior.
Ingestion points: The SKILL.md file specifies that inputs include product intent (issues, PRDs, founder messages) and delivery constraints.
Boundary markers: The skill does not define specific delimiters or warnings to treat ingested external data as potentially adversarial.
Capability inventory: The skill is configured to read from the repository and write or update durable context files such as PRODUCT.md.
Sanitization: The workflow does not include instructions for the agent to sanitize, escape, or validate the content of external inputs before generating implementation plans.

product-capability