Loop Evaluator — Development agent for Claude Code
Evaluate artifact strictly against active loop criteria and return structured pass/fail output.
How to install Loop Evaluator
Installs to ~/.claude/agents/maister-dev-maister-loop-evaluator.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/maister-dev/maister/HEAD/.claude/agents/loop-evaluator.md -o ~/.claude/agents/maister-dev-maister-loop-evaluator.md Restart Claude Code, or start a new session, for it to be picked up.
What Loop Evaluator does
name: loop-evaluator description: Evaluate artifact strictly against active loop criteria and return structured pass/fail output. Use proactively within /aif-loop when an objective verdict is needed. tools: Read, Glob, Grep model: inherit permissionMode: plan maxTurns: 6
You are evaluator.
Input:
- artifact markdown
- active rules
- optional prep results (
test-prep,perf-prep,invariant-prep)
O
Alternatives in Development
- SLO Implement — You are an SLO (Service Level Objective) expert specializing in implementing reliability standards a... - wsho 31.9k ★
- Example Curator — Use this agent when you need to evaluate CLAUDE.md examples for inclusion in the awesome-claude-md repository 570 ★
- Screen Analyst — Use this agent exclusively in the /b3:screen pipeline to fetch precomputed screening metrics for a B3 ticker 124 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Loop Refiner
Apply critique issues to produce a minimally changed improved artifact. Use proactively within /aif-loop when
Loop Invariant Prep
Prepare invariant and consistency checks for the current loop iteration. Use proactively within /aif-loop in p
Harness Evaluator
Use this agent when the harness needs to evaluate a generator's output. The evaluator interacts with the runni
Adversarial Judge.Agent
Use to adversarially evaluate a proposal or implementation against its repository profile and specs, returning
Wave Executor
Executes all tasks in a single wave of a plan. Implements, verifies acceptance criteria, and commits atomicall
Qrspi Finding Verifier
Score a single reviewer finding 0–100 against the /code-review confidence rubric. Read the per-finding file +
Related Skills
Auto Swarm Nth
Nth-iteration agent swarm — spawns parallel agent waves, evaluates strictly per wave, re-swarms gaps until 100
Claude Technique Evaluator
Evaluate new Claude prompting patterns, tools, or workflow changes. Produces go/no-go recommendations and inte
Loop Execution Evaluator
Verify implementation quality. Evaluate-Loop Step 4.
