agentsop-agent-topology-selection
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a high-level documentation and strategy layer (overlay). It does not contain any functional code (Python, JavaScript, or shell scripts) and performs no runtime operations such as network requests, file modifications, or process execution.
- [DATA_EXPOSURE]: The skill includes hardcoded absolute file paths to the author's local development environment (e.g.,
/Users/5imp1ex/Desktop/Skill-Workplace/output/...). These paths are used for documenting the source of information but do not expose sensitive system credentials or private user data. - [METADATA_POISONING]: The YAML frontmatter and README accurately reflect the skill's purpose as a topology selection rubric. The instructions are consistent with established multi-agent patterns (Supervisor, Swarm, Sequential) and do not contain deceptive directives or prompt injections.
Audit Metadata