The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest untrusted external data (SMS, social mentions, and meeting transcripts) and transform it into signals stored in the agent's persistent memory ("brain"). This creates a potential surface for indirect prompt injection attacks, where malicious instructions embedded in external payloads could be executed or followed by the agent during subsequent processing or memory retrieval.
Ingestion points: Incoming webhook payloads from external services (SMS, Twitter, etc.) as defined in SKILL.md.
Boundary markers: The skill mentions stripping HTML and escaping scripts, but it does not specify the use of clear delimiters or instructions for the agent to ignore embedded commands within the transformed text when it is later stored or read.
Capability inventory: The skill uses put_page and add_timeline_entry to write to the agent's knowledge base and triggers automated entity extraction and enrichment.
Sanitization: While the skill performs basic sanitization against traditional web attacks (XSS), it does not address the risk of natural language instructions designed to manipulate the LLM's logic.

webhook-transforms