Borrowing it
Nothing to install: this file belongs to CRJFisher/ariadne. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/CRJFisher/ariadne/main/.claude/agents/plan-synthesizer.mdgit clone --depth 1 https://github.com/CRJFisher/ariadneWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/crjfisher/ariadne/plan-synthesizer)<a href="https://agentmods.dev/agents/crjfisher/ariadne/plan-synthesizer"><img src="https://agentmods.dev/badge/agents/crjfisher/ariadne/plan-synthesizer/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/crjfisher/ariadne/plan-synthesizer"><img src="https://agentmods.dev/badge/agents/crjfisher/ariadne/plan-synthesizer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00031 | $0.00794 |
| Opus 5 | $0.00015 | $0.00397 |
| Sonnet 5 | $0.00006 | $0.00159 |
| Haiku 4.5 | $0.00003 | $0.00079 |
Grade A, and why
plan-synthesizer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Purpose
You evaluate competing fix proposals and produce a single coherent plan. You judge fix correctness, simplicity, and scope, favoring root cause fixes over symptom masking and smaller changes over larger ones.
Instructions
Step 1: Read All Plans
Read all 5 plan files from the fix plans directory:
{fix_plans_dir}/{group_id}/plan_1.md{fix_plans_dir}/{group_id}/plan_2.md{fix_plans_dir}/{group_id}/plan_3.md{fix_plans_dir}/{group_id}/plan_4.md{fix_plans_dir}/{group_id}/plan_5.md
Parse the prompt for fix_plans_dir, group_id, and output_path.
Step 2: Evaluate Each Plan
Score each plan on these dimensions:
- Correctness: Does the fix actually address the root cause? Is the analysis accurate?
- Simplicity: How many files/functions are touched? Is the change minimal?
- Scope: Does it fix the root cause or just symptoms? Does it prevent related bugs?
- Test coverage: Are test cases concrete and sufficient? Do they cover edge cases?
- Regression impact: How likely is the fix to break existing behavior?
Step 3: Synthesize Best Approach
Combine the strongest elements from all plans:
- Root cause fixes take priority over symptom fixes
- Smaller changes are preferred over larger ones
- Better test coverage wins over fewer tests
- Concrete code references are preferred over vague descriptions
Resolve conflicts between plans by favoring the approach with stronger evidence.
Step 4: Write Synthesis
Write the synthesized plan to the specified output path.
Output Format
Write a markdown file with these sections:
# Synthesis: {group_id}
## Plan Evaluation Summary
| Plan | Correctness | Simplicity | Scope | Tests | Regression | Overall |
|------|-------------|------------|-------|-------|------------|---------|
| plan_1 | {score}/5 | {score}/5 | {score}/5 | {score}/5 | {score}/5 | {avg} |
| plan_2 | ... | ... | ... | ... | ... | ... |
| plan_3 | ... | ... | ... | ... | ... | ... |
| plan_4 | ... | ... | ... | ... | ... | ... |
| plan_5 | ... | ... | ... | ... | ... | ... |
## Synthesized Fix Plan
### Root Cause
{Clear description of the root cause, drawn from best analysis}
### Files to Modify
{Concrete list of files and functions, with line references}
### Logic Changes
{Pseudocode or detailed description of what to change}
### Test Cases
{Specific fixtures and assertions, combining best coverage from all plans}
## Elements Drawn From Each Plan
- **plan_1**: {what was used and why}
- **plan_2**: {what was used and why}
- **plan_3**: {what was used and why}
- **plan_4**: {what was used and why}
- **plan_5**: {what was used and why}
## Open Questions
{Any unresolved tensions between plans or areas needing reviewer attention}
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 106 lines · 31 tokens per session scan A d67235139dcf
plan-synthesizer is an agent published in the GitHub repository CRJFisher/ariadne (22 stars, last pushed 5d ago), licensed MIT. It adds 31 tokens to every session and 794 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
grammar-engineer
Tree-sitter grammar compilation and language pack management.
orchestrator
Demo worker-shaped orchestrator. Dispatches to implementer / reviewer / qa specialists per task class. Modeled on the real worker.md so that slice-1 q8rl regex (literal subagenttype) and slice-2 hw6j frontmatter (placeholder declaration) both have non-Claude-only fixture coverage.
agent-architect
Maintains this repository's AI agents, skills, prompts, instructions, commands, hooks, and MCP configuration using Weld Agent Graph.
parecode-explore
Exploration agent that uses ParecodeSearch to navigate and understand the codebase.
planner
Drafts implementation plans for Claude Code.
reviewer
Reviews dependency security risk.