The Agent Skills Directory

[PROMPT_INJECTION]: The skill documents a framework for ingesting and processing external webhook data via an AI agent loop, which creates a surface for indirect prompt injection where data from messaging platforms could influence agent actions.\n- Ingestion points: Data enters the system through the /:platform/webhook endpoint and is persisted in the payload column of the integration_pending_tasks SQL table.\n- Boundary markers: While the framework enforces source verification (HMAC/signatures) to ensure the request is authentic, it does not specify the use of delimiters or isolation prompts to prevent the agent from executing instructions embedded in the user's message.\n- Capability inventory: The Processor component triggers a runAgentLoop(...), granting the agent full access to its tools and reasoning capabilities based on the content of the untrusted external payload.\n- Sanitization: The documentation focuses on authentication and signature verification but lacks guidance on sanitizing or escaping the content of incoming messages before they are processed by the agent.

integration-webhooks