The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and process data from external, potentially untrusted sources to provide verifiable evidence for its claims. Malicious instructions embedded in these external sources could attempt to influence the agent's behavior.
Ingestion points: Files (via observed), web search results (via web), log entries (via tool-record), and graph data (via graph-edge).
Boundary markers: The instructions do not define specific delimiters or isolation techniques for the content retrieved from external sources.
Capability inventory: The skill explicitly mentions capabilities for reading files, running shell commands, searching the web, and accessing logs.
Sanitization: There is no mention of sanitizing or validating the content retrieved from the web or local files before it is processed by the agent.

graph-mode