external-agent-skill-research

Evaluating external agent skills requires inspecting raw SKILL.md files, repository metadata, and commit history to determine Hermes applicability. This evidence-first methodology captures frontmatter and operational workflows, generating structured guidance for integration without mutating user wikis unless explicitly requested.

Evidence-first prompt analysis collects frontmatter metadata, file contents, and commit history to assess external skill quality and safety. It identifies target class, potential integration points, risks, and alignment with the Hermes thinking framework to produce actionable recommendations.

Yes, you can compare external skill implementations without modifying your wiki. The evaluation workflow assesses risks and Hermes applicability, producing decision support and recommendations for wiki ingestion only when explicitly requested by the user.

Documenting Hermes integration risks involves inspecting raw SKILL.md files and repository history to identify potential integration points and misalignments with the Hermes thinking framework. The evaluation workflow generates actionable recommendations for adoption, adaptation, or ingestion.

External skill evaluation works with GitHub repository URLs and optional subdirectory paths to initiate an automated skill assessment. It reads raw SKILL.md files and related repository metadata to determine applicability for Hermes integration.

You should avoid automated skill assessment when you need to directly mutate user wikis without an explicit request. The workflow is designed to only document Hermes applicability, risks, and recommendations for wiki ingestion when explicitly requested by the user.