CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

verification-loop

A comprehensive verification system for Claude Code sessions.

48

Quality

60%

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 ./.agents/skills/verification-loop/SKILL.md
SKILL.md
Quality
Evals
Security

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.

The body is a strong, executable verification workflow: six sequenced phases with concrete commands, explicit stop conditions, and a clear report format. Weaknesses are minor — some phases lack remediation loops, a few sections are trimmable, and it assumes a JS/Python tooling ecosystem.

DimensionReasoningScore

Conciseness

The phase commands are lean and assume Claude's competence ('npm run build 2>&1 | tail -20', 'npx tsc --noEmit'), with only minor trims needed: the body repeats the frontmatter description verbatim and the 'Continuous Mode' and Hooks sections add little. This fits 'Efficient; minor instances of over-explanation that could be trimmed' rather than 3, since the padding is a small fraction of an otherwise tight document.

4 / 5

Actionability

Concrete executable commands throughout ('pnpm build', 'ruff check .', 'npm run test -- --coverage', 'grep -rn "sk-" --include="*.ts"') plus a fill-in report template. Not a 5 because commands assume specific project tooling exists and the 'Continuous Mode' mental-checkpoint guidance is not executable; not a 3 because nothing is pseudocode.

4 / 5

Workflow Clarity

Six clearly sequenced phases with explicit checkpoints ('If build fails, STOP and fix before continuing', 'Fix critical ones before continuing'). Not a 5 because Phases 3-5 (lint, security, diff) lack remediation feedback loops when issues are found; not a 3 because key stop conditions are explicit rather than implicit.

4 / 5

Progressive Disclosure

No bundle files exist and all content (commands plus report template) appropriately lives inline in one well-organized file with clear section headers. Minor organization gaps from the tangential 'Continuous Mode' and Hooks sections, and at ~121 lines it exceeds the under-50-line simple-skill case that would default this dimension to 5.

4 / 5

Total

16

/

20

Passed

Description

25%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 too vague to function as a skill description: it names a domain but provides no concrete capabilities, no trigger phrases, and no usage guidance. A user or Claude would have difficulty knowing when to invoke it or distinguishing it from general testing/review skills.

Suggestions

List concrete actions, e.g. 'Runs builds, type checks, lint, tests with coverage, secret scans, and diff review, then produces a pass/fail verification report'

Add an explicit trigger clause, e.g. 'Use after completing a feature, refactoring, or before creating a PR when you need to verify quality gates pass'

Include natural user keywords such as 'verify', 'quality gates', 'run tests', 'pre-PR check' to improve trigger matching and distinctiveness from general code-review skills

DimensionReasoningScore

Specificity

The description names the domain ('verification system for Claude Code sessions') but lists no concrete actions — 'comprehensive' is buzzword padding with no specific verbs like 'runs builds' or 'type-checks code'. This matches the anchor 'Names the domain but actions are minimal or generic' rather than anchor 1 because a domain is named.

2 / 5

Completeness

The 'what' is vague ('comprehensive verification system' says nothing about what is actually verified) and the 'when' is entirely absent — no 'Use when...' clause, which caps completeness at 3. This fits 'Has a vague what and no when' rather than anchor 3, which requires a clear 'what'.

2 / 5

Trigger Term Quality

The only keyword present is 'verification' — a single generic term. Natural phrases users would say ('verify my changes', 'run tests before PR', 'quality gates') are entirely missing, matching 'One or two generic keywords; missing the natural phrases users say'.

2 / 5

Distinctiveness Conflict Risk

'Verification system for Claude Code sessions' is very broad and would overlap with testing, code-review, and run/launch skills. It is not anchor 1 because it is scoped to Claude Code sessions rather than being entirely generic, but it has no distinct triggers to earn anchor 3.

2 / 5

Total

8

/

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

ysyecust/everything-claude-code

Path

.agents/skills/verification-loop/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.