ln-402-task-reviewer
Pass
Audited by Gen Agent Trust Hub on Apr 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches missing instruction files, environment contracts, and reference documentation from the vendor's official GitHub repository (levnikolaevich/claude-code-skills) if local files are unavailable.
- [COMMAND_EXECUTION]: Executes shell commands via Bash to perform "Mechanical Verification" (Step 8), specifically running linting and typechecking tools discovered within the project environment.
- [PROMPT_INJECTION]: Susceptible to indirect prompt injection due to the processing of untrusted external content.
- Ingestion points: Reads Linear issue descriptions and comments, task files (e.g., TNNN-*.md), and project source code diffs.
- Boundary markers: No explicit delimiters or isolation instructions are used when interpolating task content into the agent's context.
- Capability inventory: The skill possesses significant capabilities, including executing shell commands (Bash), performing network requests (WebFetch), and modifying both local files and external issue trackers (save_issue).
- Sanitization: There is no evidence of sanitization or filtering for the content of the tasks or code being reviewed, allowing potentially malicious instructions to reach the underlying model.
Audit Metadata