CtrlK
BlogDocsLog inGet started
Tessl Logo

pr-review

仅当用户明确手动指定使用 pr-review skill 时触发;不要因为用户传入 PR 链接、要求 review 或要求代码审查而自动触发。

56

Quality

64%

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 ./.agents/skills/system/pr-review/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

77%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, highly actionable PR-review workflow with strong sequencing and an explicit validation gate. Its main weakness is progressive disclosure: the per-dimension checklists point to external files that are absent from the bundle, so the detailed review criteria are unreachable.

Suggestions

Ship the referenced checklist files (backend-quality/*.md, frontend-quality/*.md, style/*.md) inside the bundle, or inline the key acceptance criteria so the links resolve.

For the outward-facing review submission, add a brief pre-submit checkpoint (e.g. 'Review the draft report and severity counts before posting') to guard the batch comment POST.

Drop the duplicated trigger-constraint block in the body since it already lives in the frontmatter description, trimming token cost.

DimensionReasoningScore

Conciseness

The body is mostly lean — concrete gh/git commands, dense checklists, a report template — with only minor padding such as repeating the trigger constraint (line 10) already stated in frontmatter and a somewhat long template.

4 / 5

Actionability

Core mechanical steps are copy-paste ready (gh pr view/diff/checks, git worktree, the GitHub API review POST with full JSON), but the actual review criteria in phases 2-4 are delegated to referenced files that are not in the bundle, leaving gaps.

4 / 5

Workflow Clarity

A clearly sequenced staged process (Step 0 → Phase 1 → parallel Phases 2-6 → report) with an explicit phase-gate feedback loop (skip to reporting if Phase 1 is blocked), per-phase checklists, and a severity/decision taxonomy.

5 / 5

Progressive Disclosure

Structure is well organized and references are clearly signaled (checkbox + link), but the many referenced files (./backend-quality/*.md, ./frontend-quality/*.md, ./style/*.md, ../../../design/api/index.md) do not exist in the bundle, so navigation is broken.

3 / 5

Total

16

/

20

Passed

Description

51%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.

The description is highly distinctive and trigger-safe but inverts the usual structure: it is almost entirely a negative trigger constraint with no statement of capability. This makes it unlikely to fire incorrectly but leaves Claude and users guessing about what the skill actually does.

Suggestions

Lead with an explicit 'what' clause, e.g. 'Performs staged code review of pull requests: validates requirements and logic, runs parallel quality checks, then posts a review report.'

Add 2-3 concrete actions (backend/frontend quality checks, style-compliance checks, test-coverage and regression-risk assessment) so specificity rises above domain-naming.

Keep the explicit manual-trigger gate, but pair it with the positive trigger phrase 'Use when the user explicitly asks for the pr-review skill' so both what and when are present.

DimensionReasoningScore

Specificity

The description names the pr-review/代码审查 domain ("pr-review skill", "要求代码审查") but lists no concrete actions the skill performs, only a trigger constraint.

2 / 5

Completeness

It gives an extremely explicit "when" (only trigger on explicit manual naming) but never states "what" the skill does; the function must be inferred from the name, which the guidelines say not to do.

2 / 5

Trigger Term Quality

It surfaces natural user terms — "PR 链接", "review", "代码审查", "pr-review" — that users would actually say, though it misses some synonyms like "pull request" or "diff".

4 / 5

Distinctiveness Conflict Risk

The entire description is dedicated to establishing a distinct, gated trigger and explicitly suppressing overlap (do not auto-trigger on PR links or review requests), giving minimal conflict risk.

5 / 5

Total

13

/

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.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 11 missing, 2 suspicious

Warning

Total

15

/

16

Passed

Repository
labring/FastGPT
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.