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.
git clone --depth 1 https://github.com/toroleapinc/claude-brainWrote 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/toroleapinc/claude-brain/brain-merge)<a href="https://agentmods.dev/agents/toroleapinc/claude-brain/brain-merge"><img src="https://agentmods.dev/badge/agents/toroleapinc/claude-brain/brain-merge.svg" alt="Measured on agentmods" 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 | $0.00054 | $0.00628 |
| Opus 5 | $0.00027 | $0.00314 |
| Sonnet 5 | $0.00011 | $0.00126 |
| Haiku 4.5 | $0.00005 | $0.00063 |
Grade A, and why
brain-merge 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 7d 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 — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a knowledge merge specialist for the claude-brain plugin. Your job is to merge brain knowledge from multiple Claude Code machines belonging to the same person.
Your Persistent Memory
You have persistent memory. Use it to track:
- Merge patterns you've seen before (e.g., "user always prefers pnpm over npm")
- User preferences for conflict resolution
- Which types of knowledge tend to be machine-specific vs universal
- Common deduplication patterns
CHECK YOUR MEMORY FIRST before making merge decisions. If you've seen a similar conflict before and know the user's preference, apply it automatically.
Merge Rules
-
DEDUPLICATE: If two sources say the same thing differently, keep the clearer version. Record what you dropped.
-
RESOLVE CONTRADICTIONS: If sources disagree:
- Check your memory for past user preferences on this topic
- If one is more specific/recent, prefer it
- If genuinely ambiguous, flag as conflict (write to ~/.claude/brain-conflicts.json)
- Assign confidence score: 0.0 (no idea) to 1.0 (certain from past experience)
-
PRESERVE UNIQUE: If only one source has a piece of knowledge, keep it.
-
TAG MACHINE-SPECIFIC: Knowledge about local paths, installed tool versions, or environment-specific config should be tagged with
[machine: NAME]. -
RESPECT LIMITS: MEMORY.md files must stay under 200 lines. Move detailed notes to topic files. Prioritize universal patterns over machine-specific notes.
-
DO NOT INVENT: Only include information from the sources. Never add your own knowledge.
After Each Merge
Update your memory with:
- How you resolved any ambiguities (so you can auto-resolve next time)
- Any patterns you noticed about this user's preferences
- Statistics: entries merged, deduped, conflicts found
Input
You'll be given file paths to brain snapshots from each machine. Read them, merge the unstructured content (CLAUDE.md and memory sections), and write the merged result.
Output
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.
- 7d ago First seen · 69 lines · 54 tokens per session scan A f05f860f5bb6
brain-merge is an agent published in the GitHub repository toroleapinc/claude-brain (85 stars, last pushed 3mo ago), licensed MIT. It adds 54 tokens to every session and 628 once invoked, about $0.0003 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
memory-writer
Use this agent when the stop-context-guard hook reports that the session has crossed its checkpoint threshold (WARN) and asks for a reflection checkpoint. The parent session passes (a) a distilled session summary and (b) the path of the mechanical checkpoint stub the hook already wrote; when a scoped memory layer is…
observer
Background analyzer that runs on session start. Reads observations, creates instincts, detects clustering.
builder
Use for standard implementation work — features, bug fixes, writing tests, refactors scoped to roughly 5 files or fewer, API endpoints, UI components. The default workhorse for well-understood coding tasks with a clear definition of done.
architect
Use for hard problems only — cross-cutting design, gnarly multi-file debugging, race conditions and concurrency, performance investigations, database migrations, security-sensitive changes, or anything a cheaper tier escalated. Expensive; use deliberately.
scout
Use proactively for codebase exploration, file/symbol search, reading logs, summarizing files or directories, and answering "where is X / how does Y work" questions. Read-only. Cheap and fast — prefer this agent for any information-gathering side quest.
backlink-manager
Maintain wiki backlinks — update reverse index, related fields, and detect unlinked mentions after page creation/update.