blitz-reviewer

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes several local bash scripts (check_mcp.sh, check_sdk.sh, check_skills.sh, check_key.sh) to audit the local environment and installation state.
  • [EXTERNAL_DOWNLOADS]: Fetches latest version information and configuration metadata from the official Python Package Index (PyPI), the NPM registry, and the vendor's GitHub repository (api-blitz/skills).
  • [DATA_EXFILTRATION]: The check_mcp.sh script scans sensitive configuration files in the user's home directory (such as .claude.json, .cursor/mcp.json, and Claude Desktop configs) to detect the presence of the Blitz MCP server.
  • [CREDENTIALS_UNSAFE]: The check_key.sh script sends the BLITZ_API_KEY to the vendor's API endpoint (api.blitz-api.ai/v2/account/key-info) to retrieve account status and rate-limit details.
  • [PROMPT_INJECTION]: Indirect prompt injection surface exists in the code auditing workflow, which involves scanning and processing user-provided project files for API call patterns and request bodies.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 02:58 PM
Security Audit — agent-trust-hub — blitz-reviewer