Verify code or workflow changes by running tests, typecheck, and manual checks.
npx skills add https://github.com/cheafi/Trading-bot-CC --skill verify-cheafi
Please help me install this Agent Skill. Skill: verify Source: https://github.com/cheafi/Trading-bot-CC/tree/main/.github/skills/verify Command: npx skills add https://github.com/cheafi/Trading-bot-CC --skill verify-cheafi
It prevents vague “it works” updates by forcing verification steps that produce concrete evidence before declaring a change complete.
Ask the AI to verify a specific behavior change by providing the exact claim to prove and any relevant commands, tests, or build steps available in your project.
High-intent search queries and answers about installing and using this skill.
To verify a code change, you must prove the specific behavior functions as claimed by running existing automated tests, typechecks, and direct commands, falling back to manual validation only when automation is insufficient.
Evidence-first release readiness verification prevents vague completion claims by requiring concrete proof through CI checks and tests, transparently reporting passes, failures, and remaining unverified items before finalizing.
Run build verification checks in a specific order: execute existing tests first, then typecheck or build steps, then narrow direct command checks, and finally manual validation steps to produce reliable confidence.
Yes, you can use manual validation when automated tests are insufficient for CI checks, provided you define the exact behavior to prove and transparently report the manual steps alongside any automated test passes or failures.
The best way to turn ambiguous completion claims into verified outcomes is to define the exact behavior to prove and run a verification ladder of checks, reporting only verified passes, failures, and unverified items.
DethonRun full verification commands and report fresh outputs before claiming completion.
ArabelaTsoAnalyze Isabelle, Coq, and Dafny artifacts to report verified and unverified components.
JhabbigWrite a failing test before implementing code to enforce red-green-refactor discipline.
syl2042Validate onboarding readiness by running deterministic checks and LLM evaluations against AGENTS.md and PLANS.md.
