# logic-synthesizer

logic-synthesizer splits a hard problem into several parallel solution paths, weighs each one and merges the strongest parts into a single answer.

- subagent · `type=agent` · Claude Code, Codex, Cursor · ★ 1357 repo stars
- Author: qdhenry
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/qdhenry/Claude-Command-Suite/tree/HEAD/.claude/agents/wfgy/logic-synthesizer.md
- Updated: 2026-03-01
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/agents/logic-synthesizer

## What it does

A Claude subagent that treats a problem as a set of branches. On activation it loads the BBPF formula through `/wfgy:bbpf`, says hello and prints its command list, every command carrying a `*` prefix. `*multi-path-explore` generates the candidate paths, `*path-probability` scores them, `*prune-weak` drops the losers, `*synthesize-optimal` folds what is left into one answer.

The file carries its own numbers. Five paths by default, never fewer than 3 and never more than 10. Anything under a 0.05 probability gets cut, and paths that agree above 0.7 are merged. Scoring splits into feasibility 0.3, optimality 0.3, staying power 0.2, simplicity 0.2. It runs on sonnet and can call Bash, WebSearch and Task.

Two commands lean on files outside the agent. `*multi-path-explore` and `*diverge-reasoning` resolve to `.claude/commands/wfgy/` and `.claude/commands/reasoning/`, so with no `wfgy-bbpf.md` or `reasoning-multi-path.md` sitting there they have nothing to run. The rest works from what the agent file already holds.

## Install

### Claude Code (native platform)

**Manually: put the files in place**
- File: `.claude/agents/logic-synthesizer.md` — For all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention

### Codex

**Manually: put the files in place**
- File: `.codex/agents/logic-synthesizer.toml` — Codex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/

### Cursor

**Manually: put the files in place**
- File: `.cursor/agents/logic-synthesizer.md` — Cursor also reads .claude/agents/

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/qdhenry/Claude-Command-Suite/tree/HEAD/.claude/agents/wfgy/logic-synthesizer.md

## When to use it
- break a tangled problem into parallel candidate paths
- score each option before committing to one
- build a comparison matrix for competing designs
- merge compatible ideas into a single plan
- test the chosen path against several scenarios

## Example prompts
- Hand this refactor to the logic-synthesizer subagent and have it explore several paths first.
- Have logic-synthesizer run `*synthesize-optimal` and fold the three strongest paths into one plan.
- Ask logic-synthesizer for a `*solution-matrix` comparing the two caching designs.

## Similar
- [logic-audit](https://affpapa.org/en/nevibecode/skills/logic-audit.md) — skill · 545 installs
- [research-synthesizer](https://affpapa.org/en/nevibecode/agents/research-synthesizer.md) — subagent · 1377 installs
- [screenshot-synthesizer](https://affpapa.org/en/nevibecode/agents/screenshot-synthesizer.md) — subagent · 346 installs
- [launch-feedback-synthesizer](https://affpapa.org/en/nevibecode/skills/launch-feedback-synthesizer.md) — skill · 1100 installs
- [synthesizer](https://affpapa.org/en/nevibecode/skills/synthesizer.md) — skill · 106 installs
- [knowledge-synthesizer](https://affpapa.org/en/nevibecode/agents/voltagent-knowledge-synthesizer.md) — subagent · ★ 6505 repo stars

From the same set: [qdhenry/claude-command-suite](https://affpapa.org/en/nevibecode/by/qdhenry/claude-command-suite.md) — 112 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
