CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

review-local-changes

Review your local uncommitted working-tree changes (git diff plus untracked files) and return actionable improvement suggestions. Use before committing, when nothing has been pushed yet.

Quality

81%

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

The canonical home for this skill is review-local-changes in NeoLabHQ/context-engineering-kit

SKILL.md
Quality
Evals
Security

Quality

Content

73%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 highly actionable, well-sequenced orchestration workflow with exact commands, scoring rubrics, and output templates. Its weaknesses are redundancy in the filtering explanation (stated five-plus times) and the lack of verbatim prompt templates for the Phase 2 and Phase 3 agents.

Suggestions

Consolidate the min-impact/confidence filtering explanation — currently repeated in Flag Interaction, Configuration Resolution, the Phase 3 IMPORTANT bullets, Filter application order, and the concrete example — into a single canonical statement in Phase 3.

Provide verbatim prompt templates for the Phase 2 review agents and the Phase 3 confidence/impact scoring agents, mirroring the Phase 1 template.

Move the markdown/JSON output templates and the impact-level/confidence scoring tables into a references/ file (e.g., references/output-templates.md) and link to it from the body to reduce SKILL.md token load.

DimensionReasoningScore

Conciseness

The min-impact filtering logic is explained repeatedly — in Flag Interaction ("--min-impact filters which issues appear..."), Configuration Resolution, the Phase 3 intro, two consecutive IMPORTANT bullets, "Filter application order", and the concrete example — which could be consolidated into one canonical statement. It is not a 2 because the rest of the body is dense operational instruction with no padding that explains concepts Claude already knows.

3 / 5

Actionability

Concrete, executable guidance throughout: exact commands ("git status --short", "git diff --cached --stat"), a verbatim phase-1 agent prompt template, numeric confidence/impact rubrics with defined values, and complete markdown and JSON output templates. It is not a 5 because the Phase 2 review-agent instructions and the Phase 3 scoring-agent prompt are described rather than given as copy-paste templates like the Phase 1 one.

4 / 5

Workflow Clarity

A clearly sequenced three-phase workflow with explicit checkpoints and feedback loops: exit when "If there are no changes", "If there are no issues that meet this criteria, do not proceed", a verification loop where scoring agents "double checked the issue", an ordered two-step filter, and a worked example showing exactly how the filters apply. Not a 4 because validation steps are explicit at every decision point, not merely present.

5 / 5

Progressive Disclosure

No bundle files exist (references/, scripts/, assets/ are absent), and the single file is well organized with clear section headers, numbered phases, and tables, with no nested references. It is not a 5 because the ~340-line body inlines material (full JSON/markdown output templates, impact/confidence rubric tables) that could be offloaded to a one-level-deep reference file to reduce always-loaded tokens.

4 / 5

Total

16

/

20

Passed

Description

78%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 clearly and explicitly states both what the skill does and when to use it, with concrete technical specifics (git diff, untracked files, before committing). Its main weaknesses are the second-person phrasing and missing a few natural trigger synonyms like "pre-commit" or "code review".

Suggestions

Rewrite the description in third person (e.g., "Reviews local uncommitted working-tree changes (git diff plus untracked files) and returns actionable improvement suggestions.") to avoid the second-person penalty.

Add one or two natural trigger synonyms such as "pre-commit review" or "review my changes before committing" to broaden trigger-term coverage.

DimensionReasoningScore

Specificity

"Review your local uncommitted working-tree changes (git diff plus untracked files) and return actionable improvement suggestions" names the domain precisely with two concrete actions, matching the anchor for 1-2 concrete actions that are not comprehensive; the second-person phrasing ("Review your...") costs a point per the judging guidelines. It is not a 2 because the technical specifics ("git diff plus untracked files") go well beyond merely naming the domain with generic actions.

3 / 5

Completeness

It explicitly answers both what ("Review your local uncommitted working-tree changes (git diff plus untracked files) and return actionable improvement suggestions") and when ("Use before committing, when nothing has been pushed yet") with concrete trigger phrasing. It is not a 4 because the when clause is specific and concrete rather than generic or merely implied.

5 / 5

Trigger Term Quality

Natural user phrases are present — "uncommitted", "local" changes, "git diff", "untracked files", "before committing", "nothing has been pushed yet" — giving good keyword coverage. It is not a 5 because common variations such as "pre-commit review", "review my diff", or "code review" are missing.

4 / 5

Distinctiveness Conflict Risk

The niche is clearly scoped to pre-commit, unpushed local changes ("uncommitted working-tree changes", "when nothing has been pushed yet"), keeping conflict risk minor. It is not a 5 because a generic request like "review my code" could still plausibly match a closely related general code-review skill.

4 / 5

Total

16

/

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

NeoLabHQ/context-engineering-kit

Path

antigravity/skills/review-local-changes/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.