CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

harness-review

HAR: Multi-angle code, plan, scope review. Security/quality check. Trigger: review, code review, plan review, scope analysis. Do NOT load for: implementation, new features, bugfix, setup, release.

Quality

82%

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 harness-review in Chachamaru127/claude-code-harness

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.

A dense, contract-driven dispatcher body with strong workflow sequencing and a well-structured reference layer, weakened mainly by duplicated sections and some inlined detail that belongs in references. Referenced scripts are external to the bundle, slightly limiting self-contained actionability.

Suggestions

Deduplicate the two 'Output Contract' sections and consolidate repeated read-only boundary statements to recover token budget (lowest-scoring conciseness dimension).

Move the inlined governance合格ライン checklist and TeamAgent table fully into references/governance.md and references/team-debate.md, keeping only the minimum gate summary in SKILL.md.

Either bundle the referenced scripts/*.sh helpers or add a one-line note that they are provided by the harness plugin so the commands are clearly resolvable.

DimensionReasoningScore

Conciseness

The body is information-dense and assumes Claude's competence, but it carries clear redundancy — the 'Output Contract' section appears twice (lines ~18 and ~262) and the read-only boundary is restated several times — so it could be tightened despite being mostly efficient.

3 / 5

Actionability

Provides copy-paste-ready bash invocations with concrete flags, explicit literal output strings (REVIEW_AUTOSTART/AUTO_START contract lines), and a full JSON output schema; the gap is that several referenced scripts live outside this bundle, so commands are not fully self-contained.

4 / 5

Workflow Clarity

The Minimal Flow is numbered 1-6 with explicit validation checkpoints (Review Target Detection handshake, AskUserQuestion fallback, APPROVE gate checklist with all-or-fail conditions, REQUEST_CHANGES re-review conditions), giving clear sequencing with feedback loops for error recovery.

5 / 5

Progressive Disclosure

Acts as a thin dispatcher with a mode-decision table mapping each input to specific one-level-deep references (all 12 referenced files exist), but some detail (governance合格ライン checklist, TeamAgent table, Code Review Summary) is inlined rather than fully delegated, leaving minor organization gaps.

4 / 5

Total

16

/

20

Passed

Description

83%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 concise, well-scoped dispatcher description that states concrete capabilities and provides explicit positive and negative trigger guidance in third person. It would benefit only marginally from broader synonym coverage of trigger terms.

DimensionReasoningScore

Specificity

Names the review domain plus several concrete actions ('Multi-angle code, plan, scope review. Security/quality check'), listing distinct review modes rather than generic verbs; minor coverage gaps keep it below a 5.

4 / 5

Completeness

Explicitly answers both what ('Multi-angle code, plan, scope review. Security/quality check') and when ('Trigger: review, code review, plan review, scope analysis. Do NOT load for: implementation, new features, bugfix, setup, release') with concrete trigger and exclusion phrases.

5 / 5

Trigger Term Quality

Trigger terms 'review, code review, plan review, scope analysis' are natural phrases users would say, with good coverage; a few common variants (e.g. 'PR review', 'diff review') are absent so it is not a 5.

4 / 5

Distinctiveness Conflict Risk

The negative-trigger list ('Do NOT load for: implementation, new features, bugfix, setup, release') sharply bounds the niche, but 'code review' / 'review' still risk overlap with adjacent review skills, so it is not a 5.

4 / 5

Total

17

/

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

referenced_paths_exist

Referenced path issues: 6 missing

Warning

Total

15

/

16

Passed

Repository

Chachamaru127/claude-code-harness

Path

opencode/skills/harness-review/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.