Content
67%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 operational core (activation flow, coaching rules, tip and rating templates) is concrete and well-sequenced, with real, clearly signaled one-level-deep bundle files. The main defects are the duplicated README-style tail sections that restate the frontmatter and trigger list, and script/feature mentions (5-gate decision tree, 'stop with the tips') that lack operational detail in the body proper.
Suggestions
Delete the trailing '## Name', '## Description', '## Usage', and '## Features' sections that duplicate the frontmatter and the 'When to invoke this skill' section, keeping at most the '## Examples' and '## Scripts' material.
Add a rule in the 'Ongoing coaching mode' section for the 'stop with the tips' turn-off that the Features section promises, and briefly state how the rules map to the 5-gate decision tree in references/coaching-rules.md.
For each script in '## Scripts', give a one-line invocation example (e.g. how and when to run prompt_rater.py during a 'rate that prompt' request) so the bundle is actually reachable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The core coaching rules are lean and imperative, but the trailing '## Name', '## Description', '## Features', '## Usage', and '## Examples' sections (~60 lines) largely duplicate the YAML frontmatter and the earlier 'When to invoke this skill' section. This is more than the 'minor instances of over-explanation' of anchor 4, fitting anchor 3: mostly efficient but includes unnecessary content that could be trimmed. | 3 / 5 |
Actionability | The body gives copy-paste-ready templates for the tip format, prompt rating, and glossary entry format, plus an exact first-activation question and sequence. It falls short of anchor 5 because the three scripts are listed with no invocation commands, and behaviors like 'stops coaching when the user says stop with the tips' and the '5-gate decision tree' are mentioned only in the duplicated tail without operational detail. | 4 / 5 |
Workflow Clarity | The first-activation flow is clearly sequenced (Step 1 to Step 3) with an explicit skip condition ('If the user already mentioned their use case... skip this question'), and ongoing coaching rules are ordered with do/don't trigger lists. Anchor 5 is not reached because the connection between the rules and the deferred 5-gate decision tree in references/coaching-rules.md is not spelled out, and the user-initiated turn-off ('stop with the tips') is absent from the rule set. | 4 / 5 |
Progressive Disclosure | Both references exist and are one level deep with clear read conditions ('Read on first activation and consult when surfacing tips', 'Read if uncertain whether a moment is coachable'), and the three scripts are listed with one-line purposes. Minor gaps keep it from anchor 5: the scripts section gives no usage/navigation guidance, and the duplicated tail sections blur the overview's focus. | 4 / 5 |
Total | 15 / 20 Passed |