Content
92%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.
The body is an exemplary lean operator guide: concrete commands with explicit limits and defaults, strong write-path guardrails (preview/--confirm, ticket-gate, verify-before-resend), and clear routing to sibling skills. The only real gap is the underspecified generic 'call' sub-command.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is maximally dense with zero padding: no concept explanations, and every sentence carries an operating rule (limits, durations, defaults, anti-patterns like "不无限监听,不把样本当完整历史"). Matches anchor 5: 'Lean and efficient; assumes Claude's competence; every token earns its place.' | 5 / 5 |
Actionability | Commands are concrete and copy-paste ready with flags, defaults, and ranges (list --limit 20, send --conversation ... --confirm, listen --duration 1s-10m). Minor gaps keep it below 5: the generic 'tiktok-im call <resource> <action> [平台选项]' entry is underspecified and no example uses a concrete value. | 4 / 5 |
Workflow Clarity | This simple skill has unambiguous single actions plus explicit validation checkpoints and feedback loops for the risky write path: preview-by-default with --confirm, ticket-guard material blocking requests before they fire, "超时/断连先核对结果,不换通道重发" (verify results before resending), and mandatory real-account verification of responses, delivery, and permissions. The destructive/batch cap at 3 does not apply because validation is present. | 5 / 5 |
Progressive Disclosure | At 24 lines with no bundle files (references/, scripts/, assets/ absent) and no external references needed, the simple-skill exception applies: content is clearly organized (overview → command block → behavioral rules → routing to sibling skills) and easy to navigate. Matches anchor 5 under the under-50-lines guideline. | 5 / 5 |
Total | 19 / 20 Passed |