qodo-review

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides detailed shell templates for managing long-running background processes. These scripts include recipes for process backgrounding, signal handling, and status monitoring using temporary files and standard utilities like mktemp, tail, and kill.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests potentially untrusted data from the local filesystem and external URLs to provide context for code reviews.\n
  • Ingestion points: Local code diffs (including untracked files), session context JSON files, and ticket/specification URLs from services like Jira and Linear.\n
  • Boundary markers: The instructions mandate a structured JSON format for session context and provide guidelines on how to describe intent without bypassing the reviewer's judgment.\n
  • Capability inventory: The skill uses the qodo CLI to read files and perform network operations with the Qodo review engine.\n
  • Sanitization: The skill documentation notes that secrets, binary files, and gitignored files are automatically filtered and excluded from the review process to prevent accidental data exposure.\n- [EXTERNAL_DOWNLOADS]: The skill instructions include commands for updating the qodo CLI and agents, referencing official vendor update services at get.qodo.ai.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 02:06 AM
Security Audit — agent-trust-hub — qodo-review