convex-docs

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a shell command (node -p "require('./node_modules/convex/package.json').version") to identify the version of the installed Convex package. This is a benign operation used strictly for documentation version-pinning.
  • [EXTERNAL_DOWNLOADS]: The skill retrieves documentation content from docs.convex.dev. As this is the official domain belonging to the package vendor, this is a legitimate and safe operation.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it processes external documentation and search results.
  • Ingestion points: docs.convex.dev, GitHub README files for @convex-dev/* components, and web search fallback.
  • Boundary markers: The skill does not define specific delimiters for external content.
  • Capability inventory: Capabilities include reading local package files, network retrieval of documentation, and execution of a version-check command.
  • Sanitization: No explicit sanitization or filtering is described.
  • The risk is minimal as the skill prioritizes official vendor resources over unverified web content.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 03:14 PM
Security Audit — agent-trust-hub — convex-docs