ai-code-reviewer
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [CREDENTIALS_UNSAFE] (HIGH): The skill contains a string matching the pattern for a secret key (sk-abc123...) within a documentation example. While intended as a negative example, hardcoded credential patterns are flagged as high risk.
- [PROMPT_INJECTION] (HIGH): The skill is susceptible to Indirect Prompt Injection. 1. Ingestion points: The skill is designed to ingest and process untrusted AI-generated code from external sources. 2. Boundary markers: There are no instructions to use delimiters or ignore embedded instructions within the code being reviewed. 3. Capability inventory: The skill suggests executing potentially dangerous commands like 'npm test' and 'npx audit-ci' on the code being reviewed. 4. Sanitization: No sanitization or sandboxing (e.g., containers) is specified before running these commands on untrusted input.
- [COMMAND_EXECUTION] (MEDIUM): The skill recommends executing several shell commands (npx tsc, eslint, npm test) which can be exploited for Remote Code Execution (RCE) if the code under review contains malicious logic or lifecycle scripts.
- [EXTERNAL_DOWNLOADS] (LOW): The use of 'npx' implies the potential for runtime package downloads from the public npm registry.
Recommendations
- AI detected serious security threats
Audit Metadata