Reform Comparator — Research agent for Claude Code
Stage-5 comparator.
How to install Reform Comparator
Installs to ~/.claude/agents/policyengine-policyengine-claude-reform-comparator.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/PolicyEngine/policyengine-claude/HEAD/agents/reform-comparator.md -o ~/.claude/agents/policyengine-policyengine-claude-reform-comparator.md Restart Claude Code, or start a new session, for it to be picked up.
What Reform Comparator does
name: reform-comparator description: Stage-5 comparator. Takes the microsim output and the prior-scores anchor, returns PASS (within expected band, write up the result) or INVESTIGATE (mismatch, trigger calibration-diagnostics). Normalizes single-year vs 10-year scores and accounts for dataset-version drift. tools: Read, Bash, Skill model: sonnet
Reform Comparator
Stage 5 of `/analyze-policy`. Compares our microsim output to the prior-scores anchor and decides whether the result is c
Alternatives in Research
- Agent Debugger — Principal Diagnostics Engineer specializing in root cause analysis, error troubleshooting, and hotfixes 957 ★
- Fable Worker Sonnet — Fable stage worker for tasks needing real reasoning — research synthesis, nontrivial code, analysis, document 837 ★
- Echo Analyst — Use this agent BEFORE planning to surface requirement gaps, hidden assumptions, and missing acceptance criteri 523 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Prior Scores Finder
Finds prior scored reforms to anchor a new analysis — PolicyEngine's own published scores first, then official
Profile Synthesizer
Synthesizes voice profiles from self-report scores and computational writing analysis. Merges dual outputs usi
Role Prioritizer
Cheap, uncapped first-pass triage for newly-added roles. Fetches the JD, reads location, scores Priority using
Review Research
Reviews a quest's research.md before the creator is asked to accept it. Dispatched by the quest skill at revie
Gauntlet Cleaner
Stage 3 of the agent gauntlet. Runs CRAP analysis (coverage × cyclomatic complexity) and a general code review
Simulation Analyst
Lightweight single-pass analysis of a completed simulation run against Mechanics Bible ranges — spots anomalie
Related Skills
Reform Pipeline
End-to-end reform pipeline in one command — scan the news for a modelable reform in a topic area, score and ve
Estimate Before Build
Estimate-before-build skill: bounded task list, S/M/L/XL band estimates with uncertainty drivers named, scope-
Investigate Tool
Investigate why a tool is not working as expected. This command is for debugging tool behavior — wrong respons