kb-github-workflow-standards
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines core standards for agents that ingest and summarize external data from GitHub issues, pull requests, comments, and discussions. This untrusted content creates a surface for indirect prompt injection, where an attacker could embed instructions in a GitHub comment to influence the agent's behavior. While the standards include safety protocols for state-changing actions, they do not mandate data isolation or sanitization techniques.
- Ingestion points: The skill facilitates the collection of repository activity, discussion threads, and metadata from the GitHub API.
- Boundary markers: The instructions do not define the use of specific delimiters or 'ignore embedded instructions' warnings for external content.
- Capability inventory: Agents following these standards are authorized to discover repositories, cache account information, and generate workspace documents in both Markdown and HTML formats.
- Sanitization: No specific requirements for escaping, validating, or filtering external content are prescribed in the workflow standards.
Audit Metadata