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.
The body is a well-organized, appropriately terse instruction set that delegates the full methodology to HARNESS.md with a sensible local-then-network-then-condensed fallback chain. It consistently lands just below top marks: no CLI skeleton example, no failure-recovery loop after test runs, and its primary references sit outside the skill bundle.
Suggestions
Add a feedback loop to the Workflow (e.g., 'if tests fail, fix and re-run before updating README') to raise workflow_clarity from a sequence with checkpoints to one with explicit error recovery.
Include a minimal executable Click CLI snippet showing REPL-default and --json output patterns, so the Build mode is actionable even when the external HARNESS.md is unreachable.
Remove the redundant opening line 'Use this skill when the user wants Hermes Agent to act like the CLI-Anything builder' — it duplicates the frontmatter description Claude already reads.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes competence — tables, exact file paths, and packaging rules with no explanation of Click or basic CLI concepts. Minor trimmable instances remain: "Use this skill when the user wants Hermes Agent to act like the `CLI-Anything` builder" restates the frontmatter, and the fallback prose could be tightened, so it sits at anchor 4 rather than anchor 5. | 4 / 5 |
Actionability | Concrete guidance throughout: the exact harness directory tree, `find namespace_packages(include=["cli_anything.*"])`, the `cli-anything-<software>` console script, `pip install -e .` verification, and named test files (`test_core.py`, `test_full_e2e.py`). However, there is no example CLI skeleton or command template, and the core methodology is deferred to an external `HARNESS.md`, leaving gaps short of anchor 5's copy-paste-ready coverage. | 4 / 5 |
Workflow Clarity | The 7-step Workflow is a clear sequence (acquire → analyze → design → implement → TEST.md/tests/run → README → `pip install -e .`) with validation checkpoints at steps 5 and 7, and the Test mode specifies testing the installed command via subprocess. No explicit feedback loop for failed tests or installation errors exists, so it matches anchor 4 rather than anchor 5's error-recovery loops. | 4 / 5 |
Progressive Disclosure | Sections are well organized and the two references (`../cli-anything-plugin/HARNESS.md` with a clone fallback, `../registry.json`) are one level deep and clearly signaled, with bundle scripts/ appropriately reserved for installation. The key methodology reference lives outside the skill directory and cannot be guaranteed present when installed — a minor organization gap that keeps it at anchor 4 rather than anchor 5. | 4 / 5 |
Total | 16 / 20 Passed |