The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources which could contain malicious instructions designed to manipulate agent behavior. \n- Ingestion points: The skill fetches content from user-provided URLs in Step 2 and performs supplementary web searches in Step 3. \n- Boundary markers: Absent. The skill instructions do not specify the use of delimiters or 'ignore' directives to isolate external content from the agent's core instructions during processing. \n- Capability inventory: The skill has the capability to write and update multiple YAML and Markdown files in the local knowledge/sessions/ directory. \n- Sanitization: Absent. There is no mention of content validation, escaping, or filtering of the fetched data before it is structured into the profile and persisted to disk.

research