Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A tight, highly actionable instruction skill: concrete commands, a clear numbered send sequence with polling loop and error-recovery notes, and well-organized sections with no padding or concept explanations. No bundle files exist, so disclosure is judged on section organization alone.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and dense — assumes Claude's competence, explains no basic concepts, and gives every rule as a tight actionable directive (e.g. 'Use ". " or "; " instead of line breaks'). Every token earns its place. | 3 / 3 |
Actionability | Provides fully executable, copy-paste-ready commands — 'cmux send --surface surface:N "your prompt"', 'cmux send-key --surface surface:N enter', 'codex --yolo', 'claude --print --permission-mode bypassPermissions' — with no pseudocode. | 3 / 3 |
Workflow Clarity | The 'Sending prompts to a TUI agent' section is a clear numbered 1-2-3 sequence and polling is a re-check loop with explicit error-recovery guidance ('If a send failed, the cause was the escaped \"'). Not a destructive/batch operation, so no feedback-loop cap applies. | 3 / 3 |
Progressive Disclosure | Under 50 lines with no external references needed, organized into clearly signaled sections (Which agent to pick, Sending prompts, Polling, Remote VPS, The 4 agents, Driving interactive CLIs) — meeting the rubric's short-skill allowance for a 3. | 3 / 3 |
Total | 12 / 12 Passed |