CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

clarify-before-change

在修改前澄清范围、风险、假设与验收标准,避免误改。用于修改 Skill、AGENTS.md、项目规则、工作流规则、仓库结构、复杂代码或具有结构影响的文档,也用于优化 Skill、调整规则、重构、重组或需求不完整的场景。

Quality

76%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/clarify-before-change/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

63%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.

The body is a well-structured, mostly actionable instruction skill with a clear workflow, tiered risk gating, and a concrete decision-record template. Its two real defects are a broken reference to a nonexistent example file and redundant repetition of the ask-gate rule across three sections.

Suggestions

Fix or remove the dangling reference: examples/ambiguous-restructure.md does not exist in the bundle — either add the file or drop the link from the 验证 section.

State the ask-gate rule once (in 允许询问, which is declared canonical) and drop its re-statements in 工作流程 step 6 and the 证据门禁 closing line.

Add a short filled-in 决策记录 example (事实/假设/风险/最小路径/非目标/状态 with real-looking content) so the template is copy-paste actionable rather than a field list.

Replace the cross-skill governance note ("unified-taskflow、task-mode-router… 不得比本节更宽") — which Claude cannot act on from within this skill — with a one-line pointer, or move it to wherever those skills are governed.

DimensionReasoningScore

Conciseness

The body is dense and assumes competence with no concept-explaining padding, but the ask-gate rule is repeated three times ("只有缺失信息确实阻止继续时才提一个关键问题" in 工作流程, "只有答案会实质改变…才允许提问" in 允许询问, and "只有答案会改变目标…才提问" closing 证据门禁). Mostly efficient, but the redundancy could be tightened into the single canonical statement.

3 / 5

Actionability

Concrete, executable guidance throughout: a 6-step workflow, an explicit resolution chain (当前上下文 → Memory → 仓库/配置/测试 → 历史 trace → web → 可逆实验), a tiered audit/change/high-risk classification, and a fielded 决策记录 template. Falls short of 5 because there is no worked example of a filled decision record and the one referenced example is missing.

4 / 5

Workflow Clarity

The 工作流程 section gives a clearly numbered 6-step sequence, and the 验证 section provides an explicit pre-proceed checklist (目标路径、写入权限、删除/移动目标、验收条件可观察). Not 5 because there is no explicit recovery loop for when a verification item fails; not 3 because checkpoints are present and explicit rather than implicit. The destructive/batch cap does not apply since this skill only establishes change boundaries.

4 / 5

Progressive Disclosure

Sections are well-organized and the inline content is appropriately compact for a short skill, but the sole external reference — "[ambiguous-restructure.md](examples/ambiguous-restructure.md)" in 验证 — points to a file that does not exist (no examples/ directory is present in the bundle). A dangling reference breaks navigation and is worse than referencing nothing.

3 / 5

Total

14

/

20

Passed

Description

80%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 description: it explicitly answers both what and when with enumerated, natural trigger terms in the intended language. The main weakness is breadth — the "complex code / structurally-impacting docs" trigger overlaps heavily with generic code-modification skills.

Suggestions

Narrow the overlapping trigger: replace "修改…复杂代码或具有结构影响的文档" with situations specific to ambiguity (e.g., "改动会影响多个文件或规则语义、且验收标准不明"), so it does not compete with general code-editing skills.

Add the natural user phrasings that are currently missing, such as "改坏了/避免改错"、"确认改动方案"、"需求不明确", to improve trigger recall.

State the skill's concrete outputs in the description (重述目标、决策记录、单个关键问题) so the "what" matches the anchor-5 level of specificity.

DimensionReasoningScore

Specificity

The description enumerates several concrete actions — "澄清范围、风险、假设与验收标准,避免误改" — covering scope, risk, assumptions, and acceptance criteria. It falls short of 5 because "澄清" is a process verb rather than technical operations, and the actions are only four related facets of one activity.

4 / 5

Completeness

"澄清范围、风险、假设与验收标准,避免误改" explicitly states what the skill does, and "用于修改 Skill、AGENTS.md…也用于…需求不完整的场景" explicitly states when to use it with enumerated concrete trigger phrases — both halves are clearly and concretely present.

5 / 5

Trigger Term Quality

Good natural keyword coverage: "Skill、AGENTS.md、项目规则、工作流规则、仓库结构、复杂代码…重构、重组、需求不完整", including synonym pairs (重构/重组, 优化/调整). Not 5 because common user phrasings like "改坏了"、"确认改动方案"、"迁移"、"改配置" are missing.

4 / 5

Distinctiveness Conflict Risk

The pre-change-clarification niche is distinct, but the trigger "修改…复杂代码或具有结构影响的文档" is broad enough to fire on nearly any nontrivial code change, creating real overlap risk with general code-editing and planning skills. Specific artifact triggers (AGENTS.md, 工作流规则) keep it above 2.

3 / 5

Total

16

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 1 missing

Warning

Total

15

/

16

Passed

Repository

ooooooooooooooooooop/agent-tools

Path

skills/clarify-before-change/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.