CtrlK
BlogDocsLog inGet started
Tessl Logo

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.

63

Quality

75%

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

Fix and improve this skill with Tessl

tessl review fix ./codex/.codex/skills/harness-review/SKILL.md

The canonical home for this skill is harness-review in Chachamaru127/claude-code-harness

SKILL.md
Quality
Evals
Security

Quality

Content

50%

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 competently structured dispatcher with clear mode routing and well-signaled reference files, but it is undercut by missing bundle scripts (five referenced scripts/*.sh are absent), an inline verbatim JSON schema that bloats the overview, and validation logic delegated to references without inline checkpoints.

Suggestions

Either add the five referenced scripts/*.sh to the bundle (harness-review-closeout.sh, pre-review-cursor.sh, resolve-impl-backend.sh, review-ai-residuals.sh, review-weak-supervision-report.sh) or remove/replace those command references with inline guidance.

Move the ~50-line review-result.v1 JSON schema literal out of SKILL.md into a reference file and link to it, keeping only the verdict/acceptance-bar summary inline.

Inline the critical validate→fix→retry checkpoint for destructive/batch review operations (or at least one explicit validation gate) instead of fully delegating it to references/governance.md.

DimensionReasoningScore

Conciseness

The 335-line body is mostly operational, but it embeds an inline bash resolver probe, a duplicated mode column, and a ~50-line verbatim JSON schema literal that pads the context with detail Claude could read from references.

2 / 3

Actionability

It gives concrete commands (e.g. "bash scripts/harness-review-closeout.sh --dry-run --uncommitted") and literal contract strings, but all five referenced scripts/*.sh are absent from the bundle, so those commands are not executable as written.

2 / 3

Workflow Clarity

A 6-step Minimal Flow and Review Target Detection with an explicit REVIEW_AUTOSTART handshake give a clear sequence, but destructive/batch validation checkpoints are delegated to references rather than shown inline, leaving some checkpoints implicit.

2 / 3

Progressive Disclosure

References are one-level-deep and well-signaled via mode-to-reference tables, and all references/*.md exist, but five scripts/*.sh paths are referenced with no scripts/ directory present, so those navigation pointers are broken.

2 / 3

Total

8

/

12

Passed

Description

100%

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 tight, well-structured description: it states concrete capabilities, gives natural trigger terms, includes an explicit 'Use when'-style trigger clause, and adds negative triggers that sharply bound conflict risk. It avoids verbosity and over-claiming.

DimensionReasoningScore

Specificity

Lists multiple concrete review actions — "Multi-angle code, plan, scope review. Security/quality check" — naming the domain and several specific review types rather than vague abstraction.

3 / 3

Completeness

Explicitly answers both what (multi-angle review + security/quality check) and when via a literal "Trigger:" clause, satisfying the explicit-trigger requirement.

3 / 3

Trigger Term Quality

"Trigger: review, code review, plan review, scope analysis" are natural terms a user would say, with good coverage and no jargon-only phrasing.

3 / 3

Distinctiveness Conflict Risk

"Do NOT load for: implementation, new features, bugfix, setup, release" provides explicit negative triggers, giving the skill a clear niche unlikely to fire for the wrong skill.

3 / 3

Total

12

/

12

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

referenced_paths_exist

Referenced path issues: 6 missing

Warning

Total

15

/

16

Passed

Repository
Chachamaru127/claude-code-harness
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.