The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted content from external websites to determine business models and fee structures.
Ingestion points: The WebFetch tool is used in references/company-researcher.md to analyze arbitrary URLs provided by the user.
Boundary markers: The instructions lack explicit delimiters or safety warnings to treat fetched content as untrusted, which could lead the agent to follow instructions embedded in the external pages.
Capability inventory: The agent has broad read access to the project via Glob and Grep and can write implementation plans to the local filesystem (e.g., connect-recommend-plan.md).
Sanitization: No sanitization or verification of the content retrieved from WebFetch is mentioned before it is parsed into decision dimensions.
[COMMAND_EXECUTION]: The skill utilizes system tools like Grep and Glob in SKILL.md Step 2 to scan the local codebase for existing Stripe integration patterns. This represents an intended functionality for project discovery but provides the agent with broad access to read local file content.

connect-recommend