verification-before-completion

Run verification commands and confirm output before claiming success. Use when about to claim work is complete, fixed, or passing, before committing or creating PRs.

Install with Tessl CLI

npx tessl i github:secondsky/claude-skills --skill verification-before-completion

What are skills?

1.00x

Quality

83%

Does it follow best practices?

Impact

98%

1.00x

Average score across 3 eval scenarios

SKILL.md

Review

Evals

Evaluation results

100%

Fix the Broken Range Utility

Evidence-backed completion claims

Criteria

Without context

With context

Pre-fix test run

100%

Post-fix test run

100%

Bun used for tests

100%

Actual output included

100%

No unverified success claim

100%

Fix applied to source

100%

Full suite run

100%

No hedging language in claim

100%

Verification log file exists

100%

Without context: $0.4377 · 1m 7s · 22 turns · 23 in / 4,589 out tokens

With context: $0.4489 · 1m 18s · 22 turns · 22 in / 4,420 out tokens

95%

Complete the In-Memory Cache Module

Requirements checklist before declaring complete

Criteria

Without context

With context

Bun test used

100%

Bun build used

100%

Test output included

100%

Build output included

100%

Requirements list present

100%

Each requirement verified

100%

No unverified success claim

100%

Both checks before claim

40%

50%

TTL expiry implemented

100%

Size excludes expired

100%

Without context: $0.5206 · 1m 47s · 26 turns · 27 in / 5,221 out tokens

With context: $0.6640 · 1m 55s · 32 turns · 80 in / 6,359 out tokens

100%

Add a Regression Test for the Email Validator Fix

Regression test red-green cycle verification

Criteria

Without context

With context

Regression test written

100%

Bun used throughout

100%

Step 2 — pass output

100%

Revert step documented

100%

Step 4 — fail output

100%

Restore step documented

100%

Step 6 — pass output

100%

Three distinct runs

100%

No unverified completion claim

100%

Fix restored at end

100%

Without context: $0.8321 · 2m 10s · 40 turns · 40 in / 8,709 out tokens

With context: $0.7942 · 2m 3s · 40 turns · 88 in / 7,167 out tokens

Evaluated: 5 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Fix the Broken Range Utility Complete the In-Memory Cache Module Add a Regression Test for the Email Validator Fix

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.