github-v3-rest-api
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents interfaces for retrieving potentially untrusted content from GitHub, which serves as a surface for indirect prompt injection attacks. If an attacker embeds malicious instructions in GitHub data, the agent might execute them using its documented administrative capabilities.
- Ingestion points: Untrusted content enters the agent's context through operations like activity/list-notifications-for-authenticated-user and activity-list-events-for-authenticated-user.
- Boundary markers: The documentation does not provide delimiters or security instructions for the agent to distinguish between data and commands within the API responses.
- Capability inventory: The agent has high-privilege capabilities including references/operations/actions-create-or-update-repo-secret.md (secret management) and references/operations/actions-create-workflow-dispatch.md (manually triggering CI/CD pipelines).
- Sanitization: No sanitization or validation protocols for retrieved API data are described.
Audit Metadata