premature-close-checker

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs legitimate issue management tasks using established tools.
  • [COMMAND_EXECUTION]: Executes a local bash script for issue checking logic. The script path is based on the vendor's project structure ('fellowship-dev/pylot').
  • [PROMPT_INJECTION]: The skill processes external data (GitHub issue content) to identify incomplete tasks. While this represents a surface for indirect prompt injection, the risk is inherent to the skill's primary function and no exploitable pattern was identified.
  • [DATA_EXPOSURE]: Uses the GitHub CLI to view issue labels and status. These read operations are limited to the scope required for the task.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 03:02 AM
Security Audit — agent-trust-hub — premature-close-checker