research-manager

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (LOW): The skill is designed to fetch and process untrusted data from the web, which creates a surface for indirect instructions to influence the agent's planning logic.
  • Ingestion points: User-provided research topics and content retrieved from external websites via the WebFetch and WebSearch tools.
  • Boundary markers: Absent. The workflow instructions do not define clear delimiters or "ignore embedded instructions" warnings when passing data to the research-methodology skill or generating the task list.
  • Capability inventory: The skill can write to the local file system (research-tasks.md), manage persistent tasks, and trigger other skills.
  • Sanitization: The instructions do not specify any sanitization, filtering, or validation steps for content fetched from the web before it is incorporated into the project plan.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:30 PM