Content
78%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.
Highly actionable body with exact tool-call specifics, a concrete parameter table, and a worked example flow covering the full loop including clean exit. The main weakness is redundancy: the same rules are stated three times across the parameter table, Critical Rules, and Common Mistakes.
Suggestions
Merge 'Critical Rules' and 'Common Mistakes' into a single section (or drop one) — they restate the parameter table and exit conditions a third time, costing ~14 lines.
Replace the dot-graph workflow with a short numbered loop description or drop it, since the Example Flow already conveys the same sequence more concretely.
Add one line on handling a timed-out or empty listen response (e.g., re-prompt once, then confirm before exiting) to close the workflow's edge-case gap.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The parameter table, exit conditions, and example flow are efficient, but 'Critical Rules' and 'Common Mistakes' restate the same wait_for_response/skip_tts/exit-detection guidance for a third time (~14 redundant lines), and the dot graph duplicates the loop already shown in the example flow. This is more than the minor trimming of a 4, though the bulk is not padded. | 3 / 5 |
Actionability | Fully executable guidance: the exact tool name 'mcp__voicemode__converse', a parameter table with concrete values (wait_for_response=true, skip_tts=false, listen_duration_max=60), and copy-paste-ready example calls covering conversation start, mid-turn, and clean exit. Meets the 5 anchor for executable, common-case-covering commands. | 5 / 5 |
Workflow Clarity | The loop is clearly sequenced (call converse → process response → exit check → farewell with wait_for_response=false) with an explicit exit checkpoint every turn. It stays below 5 because edge cases like a timed-out or empty listen response are unaddressed; it is not a destructive/batch operation, so no validation cap applies. | 4 / 5 |
Progressive Disclosure | A self-contained skill with well-organized sections and no content that belongs in separate files; no external references are needed, satisfying the simple-skill bar for progressive disclosure. | 5 / 5 |
Total | 17 / 20 Passed |