get-qodo-rules

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFECREDENTIALS_UNSAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill implements an indirect prompt injection surface by ingesting rules from an external API and instructing the agent to treat "ERROR" level rules as non-negotiable constraints during code tasks.
  • Ingestion points: Rules are fetched from the Qodo platform API (references/pagination.md).
  • Boundary markers: Fetched rules are presented under structured headers like "# ᄒ Qodo Rules Loaded" (references/output-format.md).
  • Capability inventory: Uses Bash for environment checks and influences all subsequent code generation, modification, and review tasks.
  • Sanitization: There is no specified validation or sanitization for the rule descriptions retrieved from the remote API.
  • [CREDENTIALS_UNSAFE]: Accesses an API key stored in the ~/.qodo/config.json configuration file to authenticate requests to the Qodo platform.
  • [DATA_EXFILTRATION]: Extracts repository and organization identity from git remote URLs and transmits this metadata to an external endpoint to query for relevant coding rules.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 07:39 AM