Content
93%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.
An exemplary lean CLI skill: setup prerequisites, read and write commands, worked examples for every common operation, an honest 'delete is not supported' limitation note, and dry-run safety flags. The only refinement left is folding the dry-run preview into the modify workflow as an explicit step and adding brief error-recovery guidance.
Suggestions
Make the dry-run preview an explicit numbered step in the modify workflow (e.g., '1. get ID, 2. `--dry-run` preview, 3. apply with `--auth-token`) instead of a separate bullet.
Add one line on error recovery for update failures (invalid UUID, expired auth token) to close the feedback-loop gap.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is almost entirely executable commands and tool-specific facts (auth token, Full Disk Access, THINGSDB, dry-run semantics); it never explains what Things 3 is or concepts Claude already knows. Every line earns its place — matching the 'lean and efficient' anchor. | 5 / 5 |
Actionability | Commands are copy-paste ready and cover the common cases: `things add "Buy milk" --notes "..." --when today`, adding to a project/heading, tags, checklist items, STDIN, and the full update workflow (`things search "milk" --limit 5` → `--auth-token` → `things update --id <UUID> ...`). This matches the fully-executable anchor. | 5 / 5 |
Workflow Clarity | The modify workflow is explicitly sequenced (get the UUID first, set auth, then update) and safe-preview checkpoints exist (`things --dry-run add` / `things --dry-run update ... --completed`), so the destructive-op cap does not apply. It falls short of 5 because dry-run is presented as an aside rather than an explicit validation step in the sequence, and there is no error-recovery guidance (e.g., what to do if the UUID is wrong or the write fails). | 4 / 5 |
Progressive Disclosure | The skill is a compact, single-tool reference (~50 content lines) with no external files (references/, scripts/, assets/ are absent) and no need for them; sections (Setup, Read-only, Write, Examples, Delete, Notes) are clearly organized, matching the simple-skill guidance that well-organized sections suffice. | 5 / 5 |
Total | 19 / 20 Passed |