Content
87%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a tight, highly actionable reference with concrete commands and strong safety guidance, weakened only by the absence of an explicit error-recovery feedback loop for its destructive operations.
Suggestions
Add a short validate->fix->retry loop for destructive commands (e.g. run with --dry-run, inspect the planned mutation, then execute) to lift workflow clarity.
Show a minimal worked example of a write (e.g. trash or send) with the confirm-and-dry-run pattern applied end-to-end.
Note what to do when 'gog auth list --check' fails (re-auth step) so the safe-start sequence has an explicit recovery path.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — it never explains what Gmail or a CLI is, and every line (safe-start commands, safety bullets, command table) earns its place. | 3 / 3 |
Actionability | It provides concrete, copy-paste-ready commands with explicit flags (--readonly, --account, --json --wrap-untrusted, --dry-run) and a command table with per-command purposes, plus pointers to --help and schema for exact syntax. | 3 / 3 |
Workflow Clarity | A clear safe-start sequence and pre-flight safety checks exist (confirm account/object/mutation, --dry-run, --readonly), but the destructive operations it lists (batch, trash, delete) lack an explicit validate->fix->retry feedback loop, which caps the score at 2. | 2 / 3 |
Progressive Disclosure | The skill is under 50 lines with no bundle files and is cleanly organized into 'Safe start' and 'Commands' sections; the single off-loaded reference to ../gog/SKILL.md is clearly signaled and only one level deep. | 3 / 3 |
Total | 11 / 12 Passed |