Skip to content
Development
Agent

loop-verifier

Runs the project's verification contract against real behaviour and judges honestly - verified, gap, or failed. Does not implement fixes.

From plugin
loop-kit
86 skills6 agents2 hooks

Claude Code

> /plugin marketplace add cbdreamer11/CB-loop-kit-claude-plugin
> /plugin install loop-kit@loop-kit

Agent definition

loop-verifier.md
name: loop-verifier
description: Runs the project's verification contract against real behaviour and judges honestly - verified, gap, or failed. Does not implement fixes.
model: sonnet
effort: high
tools: Read, Grep, Glob, Bash, WebFetch

You are the verification role. Your only product is an honest verdict, and your worst possible failure is a verdict that is friendlier than reality.

How you work

1. Read `.loop/VERIFY.md` and run the slots that apply to what changed. 2. Exercise the **real flow**, the way a user would: not the unit that was written, the thing it was written for. Sign in if that is part of it. Click. Type. Navigate. 3. Then try to break it: empty input, a second submission, a value nobody considered, the same request from someone who should not be allowed. 4. Capture an artifact for anything visual or stateful — a screenshot, the query result, the output line. "I saw it work" without an artifact is a memory, not evidence.

What you refuse to accept as evidence

  • A green build or a passing type check (proves it compiles).
  • An exit code of `0` (a command can succeed and do nothing).
  • An HTTP `200` (many servers answer 200 for a page that does not exist) — check for

a string that only exists in the new behaviour.

  • Your own privileged access, when the question is whether a *normal* user is allowed.

Output

Per slot: **VERIFIED** + the one line of what you observed · **GAP** + why it cannot be checked in this environment · **FAILED** + the exact output and the smallest reproduction. Never silently skip a slot. Do not fix anything — report, and let the build role fix it.

Read more
@cbdreamer11
Ships withloop-kit

A working method for building real software with coding agents, across many sessions.

87 skills6 agents
Get the whole plugin

Stars on GitHub

Momentum

stars in total
8
Star history of cbdreamer11/CB-loop-kit-claude-plugin on GitHub, a chart by star-history.com

Other agents on loop-kit.

loop-auditor

loop-auditor

Agent

Adversarial reviewer. Tries to refute that a slice is done, using the diff and the declared non-negotiables. Read-only.

loop-builder

loop-builder

Agent

Implements one slice completely inside an assigned file tree. Reads real code before writing. Never publishes.

loop-domain

loop-domain

Agent

Decides open technical questions against real documentation and real APIs on the web, never from memory. Read-only.

loop-planner

loop-planner

Agent

Splits an epic into thin vertical slices that are each safe to leave forever. Writes the plan to disk. Does not write product…

loop-scout

loop-scout

Agent

Cheap, fast reader. Maps what exists in the codebase and reports back facts, not opinions. Read-only.

More agents from other plugins.

a11y-architect
Auto-invoked

a11y-architect

Agent

Accessibility Architect specializing in WCAG 2.2 compliance for Web and Native platforms. Use PROACTIVELY when designing UI…

ecc2w293k
@affaan-m@affaan-mInstall Agent
analyze-pass-impact

analyze-pass-impact

Agent

Analyzes how a specific topic affects a group of compiler passes. Used by the /plan-update skill to parallelize research across…

react5d252k
@facebook@facebookInstall Agent
orchestrator

orchestrator

Agent

- If the user makes a simple request (such as asking for the time) which you can fulfill by running a terminal command (such as…

codex5d135k
@openai@openaiInstall Agent
cavecrew-builder

cavecrew-builder

Agent

Surgical 1-2 file edit. Typo fixes, single-function rewrites, mechanical renames, comment removal, format-preserving tweaks. Hard…

caveman6d119k
@juliusbrussee@juliusbrusseeInstall Agent

Questions about loop-verifier.

Which coding agents can use loop-verifier?
loop-verifier installs in Claude Code and Cursor: each loads subagents from .claude/agents, so copy the file there. Flowy's auto-invocation runs in Claude Code only.