CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

settings-guide-skill

当用户问 OpenLoaf 桌面端产品内的设置、偏好、API key、模型切换、主题 / 语言 / 快捷键 / 代理 / 备份、或相关故障(key invalid、model not found)该在哪配置、怎么配置时触发。**不用于**:项目里的代码配置文件如 tsconfig / .env(→直接用 Read/Edit)、一般编程问题(→直接回答)、系统级 OS 设置。

69

Quality

87%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

82%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-organized, highly actionable internal product manual: concrete UI paths, a verbatim-error-keyed decision tree, security rules with rationale, and correct/wrong answer contrasts. Remaining gaps are minor — slight overlap between the answer-strategy and error-example sections, no recovery loop when guidance fails, and a troubleshooting tree that could be offloaded to a reference file.

Suggestions

Merge 回答策略 point 2 with the 常见错误示范 table (or drop the redundant prose) so the 'give exact paths, never say 去设置里找找' rule is stated once.

Split the 故障排查决策树 into references/troubleshooting.md, keeping SKILL.md as a configuration-map overview that links to it, which would shorten the always-loaded body.

Add a terminal fallback step to 回答策略 for when the guided setting does not fix the issue (e.g., collect the exact error message and version, then check the decision tree again or escalate).

DimensionReasoningScore

Conciseness

The body is dense, product-specific knowledge Claude could not infer (UI paths, proxy example, privacy model) with essentially no explanation of concepts Claude already knows. However, there is minor redundancy: 回答策略 point 2 ("给出具体路径,不要说'去设置里找找'") restates what the 常见错误示范 table already demonstrates, fitting the level-4 'efficient with minor trimmable instances' anchor rather than level 5.

4 / 5

Actionability

Every answer is copy-paste concrete: exact UI navigation chains ("设置 → AI 模型 → Provider 管理 → 点击对应 Provider → 编辑 API Key"), a literal proxy example ("http://127.0.0.1:7890"), a correct/wrong answer table, and a decision tree keyed on verbatim error messages. As an instruction-only skill, its guidance is fully actionable and covers the common cases, matching the level-5 anchor.

5 / 5

Workflow Clarity

The 回答策略 section gives a clear 4-step sequence (identify intent → consult map → give path / walk decision tree → API-key handling) and the 兜底规则 provides an explicit fallback for unmapped requests. No validation loop exists for the case where guided settings still fail to resolve the user's issue, which sits it at level 4 ('clear sequence, most checkpoints, minor gaps') rather than 5; the destructive/batch cap does not apply since the skill is informational.

4 / 5

Progressive Disclosure

The single file is well-sectioned with no nested or buried references, and there are no bundle files, so nothing is misfiled. At ~90 lines the body exceeds the under-50-line simple-skill exception, and the troubleshooting tree covering six product areas (邮件/日历、终端、编辑器、文件管理器、界面) is a plausible candidate for an on-demand references/ file, placing it at level 4 ('good structure, most content appropriately placed') rather than 5.

4 / 5

Total

17

/

20

Passed

Description

86%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, tightly scoped description with explicit trigger conditions, natural user vocabulary including verbatim error strings, and clear exclusion boundaries. Its only weakness is that the skill's function is implied through the user's question rather than stated directly, and a few configurable areas from the body are unmentioned.

DimensionReasoningScore

Specificity

The description names the concrete domain ("OpenLoaf 桌面端产品内的设置") and enumerates specific sub-topics ("API key、模型切换、主题 / 语言 / 快捷键 / 代理 / 备份" plus exact errors "key invalid、model not found") and the action (where/how to configure). It stops short of comprehensive coverage — body-covered areas like 数据导出、存储空间、项目/Git 设置 are absent — matching the 'several specific actions with minor gaps' anchor rather than the level-5 comprehensive one.

4 / 5

Completeness

The 'when' is explicit and detailed (trigger conditions plus an explicit **不用于** exclusion list), and the 'what' (telling the user where and how to configure, "该在哪配置、怎么配置") is present but phrased as the user's question rather than the skill's function. Both exist, so it is above level 3, but the 'what' is not stated as a capability, keeping it below level 5.

4 / 5

Trigger Term Quality

It covers natural user phrasing comprehensively: 设置/偏好 (synonym pair), API key, 模型切换, 主题, 语言, 快捷键, 代理, 备份, and the literal error strings users would paste ("key invalid"、"model not found"), plus exclusion terms like tsconfig/.env. This matches the level-5 anchor of comprehensive natural-term coverage including synonyms.

5 / 5

Distinctiveness Conflict Risk

It is scoped to one named product (OpenLoaf desktop) and explicitly disambiguates adjacent territories — project code config files (tsconfig/.env → Read/Edit), general programming questions, and OS-level settings — giving it a clear niche with minimal wrong-skill trigger risk.

5 / 5

Total

18

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

itsablabla/openloaf-web

Path

apps/server/src/ai/builtin-skills/settings-guide/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.