cohesion-over-testability

Collapse test-shaped production boundaries while preserving behavior and coverage. Use when a helper, wrapper, injected dependency, or export exists mainly to let a unit test reach internals.

Quality

84%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

75%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 well-crafted, opinionated skill body: executable greps, a concrete procedure with stop conditions and counter-indications, and disciplined progressive disclosure to a real reference file. The main gaps are the undemonstrated inline edit, a missing post-change verification step for a test-deleting workflow, and duplicated audit-grep content between the body and the reference file.

Suggestions

Add a step 7 to The Procedure: after inlining, run the remaining test suite (or build) to confirm behavior is preserved — the workflow deletes tests and files but currently has no verification of the 'preserving behavior and coverage' promise.

Deduplicate the audit greps: keep only the trigger heuristics in the body's 'Audit Sweep' section and let sweep-procedure.md own the full grep scripts, since the same three greps currently appear in both.

Show the inline as a small before/after code snippet (outer + inner → single function) so the central edit is executable rather than described, rather than relying on the repo-specific @epicenter/svelte worked example.

Dimension	Reasoning	Score
Conciseness	The body is dense and assumes competence — no space is spent explaining testing concepts Claude already knows, and every section carries operational content. Minor trimmable spots remain: the three-paragraph philosophical opener, the 'Related skills' block, and the closing question 'without the test, would I have written this as two pieces?' restated from the Signal-vs-Reason section. Efficient with minor over-explanation fits 4 rather than the fully-lean 5.	4 / 5
Actionability	The audit-sweep greps are copy-paste-ready shell, the six-step procedure is concrete (caller counting, LOC ratio, inline, three payoff options), and the five smell forms each end in a specific instruction. Gaps keep it from 5: the central 'Inline' step is never demonstrated as an actual edit, and the worked example is repo-specific (commit 'd5b61aed8' of '@epicenter/svelte') rather than reproducible.	4 / 5
Workflow Clarity	The Procedure is a clearly sequenced six steps with explicit stop conditions ('If it's two or more... stop and reassess') and a 'When NOT to Inline' guard checklist that functions as pre-flight validation. Not 5: the workflow deletes tests and code yet has no post-inline verification step (e.g., run the remaining suite or confirm behavior preserved), a minor but real validation gap.	4 / 5
Progressive Disclosure	A task-based 'References' section ('For running a systematic audit across the codebase... read references/sweep-procedure.md') points to a real, one-level-deep bundle file, and the body is well-sectioned. Not 5: the body's 'Audit Sweep' section duplicates the greps already in sweep-procedure.md's Step 1, inlining content that clearly belongs in the separate file.	4 / 5
	Total	16 / 20 Passed

Description

87%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: explicit what-and-when with concrete, natural trigger phrases covering the main forms of the smell. The only weakness is a slightly thin action list and missing common synonyms (mock, seam, test-only) that some users would naturally say.

Dimension	Reasoning	Score
Specificity	Names the concrete actions ('Collapse test-shaped production boundaries while preserving behavior and coverage') and enumerates the specific shapes ('helper, wrapper, injected dependency, or export'), though the payoff options (integration test, type invariant, test deletion) are absent. It lists several specific actions with minor coverage gaps, fitting the 4 anchor better than the 1-2-action 3 anchor and short of the comprehensive 5.	4 / 5
Completeness	Explicitly answers both: 'what' ('Collapse test-shaped production boundaries while preserving behavior and coverage') and 'when' ('Use when a helper, wrapper, injected dependency, or export exists mainly to let a unit test reach internals') with concrete trigger phrases. This matches the 5 anchor exactly; the 4 anchor requires a 'when' that is only less explicit.	5 / 5
Trigger Term Quality	'helper, wrapper, injected dependency, export', and 'unit test reach internals' are natural phrases a developer would use, but common synonyms like 'mock', 'seam', 'test-only', or 'DI' are missing. Good keyword coverage with a few natural terms missing matches 4; not 5 because the synonym set is incomplete.	4 / 5
Distinctiveness Conflict Risk	A clear niche — test-shaped production splits — with triggers tied to a specific code smell, minimal overlap with generic refactoring or testing skills. Distinct triggers and niche clarity match the 5 anchor.	5 / 5
	Total	18 / 20 Passed

Validation

93%

Warnings & errors only

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

Criteria	Description	Result
relative_links	Relative link issues: 5 suspicious	Warning

	Total	15 / 16 Passed

Repository: EpicenterHQ/epicenter
Path: .agents/skills/cohesion-over-testability/SKILL.md
Commit: f9441c8

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

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.