github-project

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses standard GitHub CLI (gh) commands to interact with project items and view repository data. It also references a vendor-specific setup command, /setup-ouyangjiahong-skills, for initial configuration when missing.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources (GitHub issue/PR content) and local configuration files, which is a standard operational surface.
  • Ingestion points: Data enters through command outputs from gh pr view, gh issue view, and gh project item-list, as well as the local file docs/agents/issue-tracker.md.
  • Boundary markers: The instructions do not define specific delimiters to isolate external data from the agent's logic.
  • Capability inventory: The skill is restricted to GitHub management tasks via gh commands and has disable-model-invocation set to true, limiting autonomous reasoning risks.
  • Sanitization: No specific filtering or sanitization of GitHub metadata is mentioned, relying on the agent's standard parsing of command output.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 02:04 AM
Security Audit — agent-trust-hub — github-project