skills/metroscope-dev/skills/research/Gen Agent Trust Hub

research

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to fetch and analyze content from external primary sources such as documentation and source code, which introduces a vulnerability to instructions embedded in that external data.
  • Ingestion points: External data from websites, code repositories, and API endpoints is ingested into the agent context during the research process specified in SKILL.md.
  • Boundary markers: The instructions lack requirements for using delimiters or explicit "ignore embedded instructions" warnings to isolate untrusted external content from the agent's logic.
  • Capability inventory: The skill is authorized to write data into the local repository as Markdown files, meaning successfully injected instructions could influence how files are formatted or what content is stored.
  • Sanitization: No sanitization, validation, or escaping mechanisms are defined to clean the content retrieved from external sources before it is processed or written to disk.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 07:17 AM
Security Audit — agent-trust-hub — research