context-hub

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the chub CLI tool (located at /Users/ai/.nvm/versions/node/v24.13.0/bin/chub) to perform documentation management tasks such as searching, fetching, and annotating.
  • [EXTERNAL_DOWNLOADS]: The chub get command retrieves documentation content from an external repository managed by the chub service. This is the primary intended function of the skill.
  • [DATA_EXFILTRATION]: Commands such as chub feedback and chub annotate transmit user-provided feedback and notes to the external service. These operations are transparently documented as features for persisting and sharing documentation 'gotchas'.
  • [PROMPT_INJECTION]: The skill was evaluated for indirect prompt injection risks associated with processing external documentation.
  • Ingestion points: External documentation fetched via chub get (SKILL.md).
  • Boundary markers: None identified; documentation content is processed directly.
  • Capability inventory: The agent is instructed to use the fetched documentation to write code.
  • Sanitization: No specific sanitization or instruction-filtering is applied to the retrieved content.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 05:35 AM
Security Audit — agent-trust-hub — context-hub