typescript-review
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a pure reasoning-based extension that provides guidelines for manual code review. It does not require any additional permissions or tool access.
- [COMMAND_EXECUTION]: To mitigate risks from untrusted code, the skill explicitly commands the agent: 'Do not run tsc or any commands.' This prevents the execution of potentially malicious code snippets during the review process.
- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it reads user-provided TypeScript files. However, the risk is assessed as safe because the execution environment is highly restricted ('disable-model-invocation: true') and the skill lacks any capabilities (such as file-write or network-send) that could be triggered by an injection.
- [DATA_EXFILTRATION]: No network operations or patterns involving sensitive data harvesting were detected.
Audit Metadata