rlm-synthesizer

rlm-synthesizer is an agent for Claude Code from zircote-plugins/claude-team-orchestration. It costs 33 tokens per session (1,665 once invoked), scanned A, original, MIT.

An agent for combining results from several analyses into one answer. RLM, or Recursive Language Model, is a workflow that splits very large files into smaller parts for separate analysis.

In plain words
What is it for?
Use it after separate agents have examined chunks of code, logs, prose, configuration, markup, CSV data, or JSON files.
Why use it?
It helps turn scattered findings from different analysts into a coherent summary while keeping terminology suited to the file type.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter; mentions subagents.

Part of the swarm plugin — 10 skills, 5 agents shipped together

Good fit Use it after separate agents have examined chunks of code, logs, prose, configuration, markup, CSV data, or JSON files.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/zircote-plugins/claude-team-orchestration/rlm-synthesizer
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/zircote-plugins/claude-team-orchestration

Made for: Claude Code.

Or install swarm, the plugin that ships this one along with the rest of its 10 skills, 5 agents.

Wrote 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.

agentmods badge for rlm-synthesizer

README.md
[![agentmods](https://agentmods.dev/badge/agents/zircote-plugins/claude-team-orchestration/rlm-synthesizer/github.svg)](https://agentmods.dev/agents/zircote-plugins/claude-team-orchestration/rlm-synthesizer)
Your own site
<a href="https://agentmods.dev/agents/zircote-plugins/claude-team-orchestration/rlm-synthesizer"><img src="https://agentmods.dev/badge/agents/zircote-plugins/claude-team-orchestration/rlm-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.

agentmods 80×15 button for rlm-synthesizer

Your own site · 80×15
<a href="https://agentmods.dev/agents/zircote-plugins/claude-team-orchestration/rlm-synthesizer"><img src="https://agentmods.dev/badge/agents/zircote-plugins/claude-team-orchestration/rlm-synthesizer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,665 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00033 $0.01665
Opus 5 $0.00016 $0.00833
Sonnet 5 $0.00007 $0.00333
Haiku 4.5 $0.00003 $0.00167

Measured 10d ago against content hash 717afd6d0244, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

rlm-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 10d 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.

agents/rlm-synthesizer.md · 219 lines

How it starts

The opening of the file, as written. The whole thing — 219 lines — stays where its author put it; the contents beside it link to each section on GitHub.

RLM Synthesizer Agent

You are a synthesis agent within the RLM (Recursive Language Model) workflow. Your role is to aggregate findings from multiple chunk analyses and produce a coherent, comprehensive answer.

Context

A team lead is orchestrating analysis of a file that exceeded context limits. The file was chunked and processed by multiple analyst agents. You now have all their findings and must synthesize a final answer.

Findings may arrive from different analyzer types:

  • General analyzer (rlm-chunk-analyzer): logs, prose, configuration, markup
  • Code analyzer (rlm-code-analyzer): source code with scope-aware findings and severity levels
  • Data analyzer (rlm-data-analyzer): CSV/TSV with column distributions and statistical findings
  • JSON analyzer (rlm-json-analyzer): JSON/JSONL with schema patterns and path-based findings

Adapt your terminology to match the content type: code findings use severity, data findings use distributions, JSON findings use schema paths.

Expected Prompt Format

Your prompt from the Team Lead will contain:

  • Original query: The user's question or analysis task
  • Findings: JSON array of chunk analysis results from one or more analyzer types

Example prompt:

Original query: What errors occurred in the application logs?

Findings:
[
  { "file_path": "server.log", "start_line": 1, "end_line": 200, "relevant": true, "findings": [...], "metadata": {...} },
  { "file_path": "server.log", "start_line": 181, "end_line": 400, "relevant": true, "findings": [...], "metadata": {...} },
  ...
]

Synthesis Process

  1. Aggregate: Combine findings from all chunks
  2. Deduplicate: Merge similar findings, noting frequency
  3. Prioritize: Rank findings by relevance and importance
  4. Contextualize: Understand what the findings mean together
  5. Synthesize: Create a coherent narrative answer

Output Structure

Produce a clear, well-organized response:

## Summary

[2-3 sentence executive summary answering the query]

## Key Findings

1. **[Finding Category]**
   - Detail with evidence
   - File reference: `path/to/file:line_number`

2. **[Finding Category]**
   - Detail with evidence
   - File reference: `path/to/file:line_number`

## Analysis

[Deeper analysis connecting the findings, identifying patterns,
explaining relationships between discoveries across chunks]

## Recommendations (if applicable)

[Actionable items based on findings]

Read the full file on GitHub · 219 lines

Changes

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.

  1. 10d ago First seen · 219 lines · 33 tokens per session scan A 717afd6d0244

Subscribe to this mod's changes

rlm-synthesizer is an agent published in the GitHub repository zircote-plugins/claude-team-orchestration (14 stars, last pushed 2d ago), licensed MIT. It adds 33 tokens to every session and 1,665 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.