manager

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh), git, and jq to perform project management tasks such as searching, viewing, and editing issues, as well as managing labels and project boards. These operations are core to its documented functionality and use well-known established tools.
  • [SAFE]: The skill incorporates security-conscious logic, specifically the 'public-repo gate' described in SKILL.md, which instructs the agent to sanitize or remove private file paths, CRM identifiers, and personal information before updating issues in public repositories.
  • [SAFE]: Analysis of indirect prompt injection surfaces (Category 8):
  • Ingestion points: Data is ingested from GitHub issue titles, bodies, and project fields via gh search and gh issue view (SKILL.md).
  • Boundary markers: There are no explicit instructions for using delimiters to isolate external issue content from the agent's primary instructions (absent).
  • Capability inventory: The agent can create/edit issues and labels, move project items between lanes, and read git logs/status across multiple repositories (SKILL.md).
  • Sanitization: The skill includes explicit instructions for data sanitization and verification when interacting with public repositories, mitigating accidental data leakage and ensuring verifiable completion criteria (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 08:27 AM
Security Audit — agent-trust-hub — manager