Content
87%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The content is lean, highly actionable, and well-structured with real bundled helper scripts; its only gap is that workflows are presented more as a command catalog than explicitly sequenced processes with feedback loops.
Suggestions
For destructive or interactive actions (e.g. send-keys, kill-session), add a brief validate-then-act checkpoint workflow with retry guidance to strengthen workflow clarity.
Consider trimming the PilotDeck Migration Note (a long temp path) from the skill body or relocating it to keep the overview token-efficient.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is almost entirely executable commands with minimal prose and no explanation of what tmux is, matching the 'lean and efficient; every token earns its place' anchor (the brief PilotDeck migration note is the only soft exception). | 3 / 3 |
Actionability | It provides concrete, copy-paste-ready tmux commands (e.g. 'tmux capture-pane -t shared:0.0 -p', 'tmux send-keys -t shared:0.0 -l -- "Please continue"'), matching the 'fully executable code/commands' anchor. | 3 / 3 |
Workflow Clarity | The prompt-checks section has a sequence (capture → grep → approve only when understood) with one checkpoint, but most sections are a reference catalog rather than sequenced multi-step workflows with explicit feedback loops. | 2 / 3 |
Progressive Disclosure | SKILL.md is a concise overview and points to two real, one-level-deep helper scripts (scripts/find-sessions.sh, scripts/wait-for-text.sh) clearly signaled under a Helpers section, matching the 'clear overview with well-signaled one-level-deep references' anchor. | 3 / 3 |
Total | 11 / 12 Passed |