audit-gate-logic
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data including CI configuration files (e.g., GitHub Actions YAML), repository branch protection settings, and pipeline execution logs. This creates an attack surface for indirect prompt injection where malicious instructions could be embedded in the audited files to influence the agent's findings.
- Ingestion points: The skill reads file content from
.github/workflows/*and fetches repository metadata using thegh apitool. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the agent processes external data.
- Capability inventory: The skill utilizes shell command execution via the GitHub CLI (
gh) to read repository settings. While instructions emphasize a 'read-only' audit, the agent has the capability to interact with repository APIs. - Sanitization: There are no explicit instructions for the agent to sanitize or validate the integrity of the external data before processing it for analysis.
- [COMMAND_EXECUTION]: The skill instructs the agent to use the GitHub CLI (
gh api) to probe repository settings. While this is a standard and expected tool for CI/CD auditing, it represents a capability that could be misused if the agent's context is compromised by injection.
Audit Metadata