Content
75%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A highly actionable, well-structured skill with concrete executable commands for every source and sensible safety gates, resulting in strong workflow clarity. The main costs are token efficiency — duplicated source lists and two near-identical worked examples — and missed opportunities to move per-registry detail into separate reference files.
Suggestions
Consolidate the two 'Complete Example' walkthroughs into one, or compress the second to a one-line variant showing only the ClawHub install command — the ask/review/invoke steps are already demonstrated.
Deduplicate the trusted-source repositories: list them once (in SAFETY or section 4) and cross-reference from the other spot.
Add a brief error-recovery note for the install workflow (e.g., slug not found, prefix required for ClawHub, retry with `--force` on name collision) to close the validation gap.
Move the Hermes hub source table and per-registry search detail into a references file (e.g., references/registries.md) to slim the main SKILL.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly commands and specifics, but contains real redundancy: the trusted-source repos are listed in the SAFETY section and again in section 4, and two structurally identical "Complete Example" walkthroughs repeat the same ask/install/review steps. It could be tightened into one example without losing information, matching 'mostly efficient but could be tightened' rather than the leaner 4. | 3 / 5 |
Actionability | Guidance is copy-paste ready throughout: concrete search commands (`hermes skills search <query>`, `clawhub search "calendar"`, `curl -s "https://clawhub.ai/api/v1/skills?q=calendar"`), all install forms with prefixes and version pinning, a flags table, and manual-install commands. These cover the common cases fully. | 5 / 5 |
Workflow Clarity | The Search Strategy gives an ordered 1-4 sequence and the worked examples lay out a clear 6-step flow with checkpoints (ask user, `hermes skills inspect` before installing, read SKILL.md for compatibility before invoking). It is not a 5 because there is no error-recovery guidance (e.g., what to do when an install fails or a slug does not resolve), only the one footnote about the bare-slug ClawHub case. | 4 / 5 |
Progressive Disclosure | The single SKILL.md is well organized with clear section headers (Safety, Sources, CLI reference, Locations, Search Strategy, Examples) and no nested or buried references. It falls short of 5 because at ~280 lines the per-source detail (Hermes hub source table, ClawHub specifics) would benefit from being split into one-level-deep reference files. | 4 / 5 |
Total | 16 / 20 Passed |