What problem does it solve? Writing SKILL.md files that pass the Hermes skill validator and match peer conventions is error-prone: frontmatter must start at byte zero, descriptions are truncated at 57 characters in the system prompt index, and the wrong creation tool writes to the user-local tree instead of the repo. This Skill encodes the exact validator rules, size limits, and peer-matched structure so authored skills load correctly on the first try. ## Core Features & Use Cases - Validator-Accurate Rules: Documents the hard constraints from the skill manager validator, including the 1024-character description limit, 100,000-character file cap, and byte-zero frontmatter requirement. - Peer-Matched Structure and Quality Checks: Provides the standard section layout (Overview, When to Use, Common Pitfalls, Verification Checklist) plus eight writing-quality principles to avoid no-op prose, duplication, and premature completion. - Correct Tooling Workflow: Clarifies when to use write_file versus skill_manage for creating, patching, and rewriting in-repo skills, including the session-cache caveat for newly added skills. - Use Case: When asked to add a reusable workflow skill to the hermes-agent repository, follow this Skill to survey peers, draft the SKILL.md, validate it with a Python snippet, and commit it on the active branch. ## Quick Start Create a new in-repo skill named my-workflow under the software-development category following the hermes-agent-skill-authoring conventions and validate its frontmatter.
