Skip to main content
subagent CX Codex

okf-comparison-reader — Claude Code subagent: install

The okf-comparison-reader subagent for Claude Code, Codex, and Cursor checks task answers against references. It evaluates queries, rules, numbers, and answers stating that the bundle contains no information.

no license ★ 105 updated 2026-10-31

Install into Claude Code · Manually: put the files in place All methods and platforms ↓
.claude/agents/okf-comparison-reader.md
Open on GitHub

What okf-comparison-reader does

okf-comparison-reader issues a verdict for each task by comparing its question, reference answer, and derived answer. It receives no full bundle or producer assessment. Its available tools are read, grep, find, ls, and bash.

For query tasks, it compares result rows in light of tables, joins, filters, and parameters. If the database can be reached, it runs both queries and compares their output. Otherwise, it reviews the query text and says the result needs manual confirmation.

What it helps with

  • compare each task answer with its reference
  • check whether two queries return matching rows
  • judge answers about rules and numbers

Example prompts for the agent

  • Compare the task answers with their references and give a verdict for each one.
  • Check whether the two queries in this task return the same rows.
  • Judge this answer about a rule against its reference.

How to install okf-comparison-reader

Manually: put the files in place

.claude/agents/okf-comparison-reader.md

For all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention

Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean

Questions about okf-comparison-reader

How do I install okf-comparison-reader in Claude Code?

Save the subagent file at .claude/agents/okf-comparison-reader.md.

What does okf-comparison-reader do?

It compares each derived task answer with its reference and gives a verdict. For query tasks, it checks the returned rows and runs both queries when the database is reachable.

How does it assess queries when the database is unreachable?

It reviews the query text and notes that the result needs manual confirmation.

Similar subagents

All subagents for Testing and QA →

From the same set8

okf-writer The okf-writer Claude Code subagent drafts concepts from source records under a run’s rules. It handles one part of a bulk produced with okf-produce. CX Codex ★ 105repo stars okf-runner The okf-runner Claude Code subagent runs one computation or read-only check from a brief. It returns a run record with the tool output and any error. CX Codex ★ 105repo stars okf-derivation-reader The okf-derivation-reader Claude Code subagent answers a task set using an OKF bundle. It works from the bundle alone, without task references or extra context. CX Codex ★ 105repo stars okf-faithfulness-reader The okf-faithfulness-reader Claude Code subagent checks concept claims against the passages cited in their footnotes. It flags unsupported claims, conflicting facts, and questions the sources leave open. CX Codex ★ 105repo stars okf-validate The okf-validate Claude Code skill checks an Open Knowledge Format (OKF) bundle against version 0.2 of its specification and explains each finding. It runs when a bundle is reviewed or after a producer creates or updates one. CX Codex +1 ★ 105repo stars okf-answer The okf-answer Claude Code skill answers questions using Open Knowledge Format bundles. It can also run bundle computations and save answers on request. CX Codex +1 ★ 105repo stars okf-produce The okf-produce Claude Code skill turns readable source material into an Open Knowledge Format (OKF) bundle and records its context. It activates for requests to create or update a knowledge base, even when OKF is not named. CX Codex +1 ★ 105repo stars

Project sponsors

Companies that keep this analytics open