agentic-workflow-orchestration
Pass
Audited by Gen Agent Trust Hub on Apr 3, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists of instructional markdown content and does not include any executable scripts, binaries, or automated code execution patterns.
- [SAFE]: External links point to the author's own GitHub repositories (hack23) or official GitHub documentation, which are considered trusted sources within the scope of this analysis.
- [PROMPT_INJECTION]: The orchestration patterns described involve agents processing potentially untrusted data from GitHub issues and comments. The skill mitigates this surface by mandating security best practices in its rules. * Ingestion points: GitHub issue titles, bodies, and comments are identified as data sources in several examples. * Boundary markers: Present. The rules mandate the use of structured data formats like JSON and YAML for communication. * Capability inventory: Examples describe agents capable of creating issues, comments, and pull requests via the GitHub tool. * Sanitization: Present. The skill explicitly requires validation of all inter-workflow messages and forbids trusting unchecked data from other workflows.
Audit Metadata