verify-with-docs

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface by instructing the agent to ingest and process untrusted external data from library documentation websites and repository source code (e.g., via opencode.jsonc references or repositories such as anomalyco/opencode).\n
  • Ingestion points: SKILL.md guides the agent to read from project manifest files (package.json, pyproject.toml, go.mod), library source code in node_modules or local clones, and external documentation URLs.\n
  • Boundary markers: The skill suggests extracting minimal relevant lines and quoting facts with citations, which acts as a manual filter but does not constitute a robust security boundary against adversarial content.\n
  • Capability inventory: The instructions require the agent to perform file-system reads and network fetches to access the external data.\n
  • Sanitization: No specific sanitization, escaping, or strict schema validation is prescribed for the content retrieved from external documentation or source repositories before it enters the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 04:40 PM