pr-autoloop
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the GitHub CLI (
gh) to gather PR state, check run statuses, and interact with the GitHub API for reactions and comments. These are standard operations for a PR maintenance tool and are performed on the repository context. - [REMOTE_CODE_EXECUTION]: The skill mentions external tools like 'chatgpt-codex-connector[bot]' but only interacts with it via standard API responses and reactions, not by downloading or executing arbitrary code from it.
- [PRIVILEGE_ESCALATION]: No sudo, chmod, or administrative privilege escalation patterns were detected. The wait mechanisms (like ScheduleWakeup or bounded bash loops) are standard for agent pacing.
- [DATA_EXFILTRATION]: The skill reads data from GitHub to process CI results. While it interacts with a network resource (GitHub API), it does so to perform its stated purpose of PR maintenance within the user's configured repository environment, not for exfiltrating sensitive local files.
Audit Metadata