Verify

SkillMonitoring & ops

Verify that a code change works through the real running application and observable acceptance criteria. Use when the user invokes $verify or asks for end-to-end confirmation beyond unit tests, lint, or type checks.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Verify skill

What this skill tells your AI

The instructions your AI receives, as published by solvcon/solvcon in .agents/skills/verify/SKILL.md and read by ahel’s review.

Translate the request and diff into concrete observable acceptance criteria. Read repository instructions and a matching run-* skill if one exists.

Build and launch the real application using the supported workflow. Exercise the changed behavior through its actual user or integration surface and collect evidence for every acceptance criterion. Include an important negative or boundary case when it is safe and relevant.

Tests, lint, logs, and internal state may support the result, but they do not replace direct observation of the running application. Do not claim success when an environment limitation prevented the relevant behavior from running.

Stop processes and remove temporary state created for verification. Report each criterion as passed, failed, or blocked, with the evidence and commands used. Do not fix a failure unless the user also requested implementation.

Signals

GitHub stars
76
Forks
73
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
verify-solvcon
Source
github.com/solvcon/solvcon