mem-quartermaster

mem-quartermaster is an agent for Claude Code from JosephOIbrahim/Synapse. It costs 104 tokens per session (777 once invoked), scanned A, original, MIT.

A coordinating agent for a multi-part MEMORY board task. It calculates shared facts once at the common starting point of the work, then gives those facts to the other agents.

In plain words
What is it for?
Use it to record the test-suite baseline, key contracts, fixed interfaces, and worktree readiness in a supply packet.
Why use it?
It prevents each agent from repeating costly checks and accidentally using a lower test standard from its own branch.

Agent for Claude Code

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.

agentmods
npx agentmods add agents/josephoibrahim/synapse/mem-quartermaster
Clone the repo
git clone --depth 1 https://github.com/JosephOIbrahim/Synapse

Made for: Claude Code.

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 mem-quartermaster

README.md
[![agentmods](https://agentmods.dev/badge/agents/josephoibrahim/synapse/mem-quartermaster.svg)](https://agentmods.dev/agents/josephoibrahim/synapse/mem-quartermaster)
Your own site
<a href="https://agentmods.dev/agents/josephoibrahim/synapse/mem-quartermaster"><img src="https://agentmods.dev/badge/agents/josephoibrahim/synapse/mem-quartermaster.svg" alt="Measured on agentmods" height="20"></a>
Per session 104 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 777 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00104 $0.00777
Opus 5 $0.00052 $0.00388
Sonnet 5 $0.00021 $0.00155
Haiku 4.5 $0.00010 $0.00078

Measured 4d ago against content hash 4651e8daf433, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mem-quartermaster 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 4d 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.

.claude/agents/mem-quartermaster.md · 66 lines

How it starts

The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are QUARTERMASTER. You put the right facts in front of a leg before it needs them, so no leg spends its budget re-deriving what is already knowable.

AGENTS.md binds you in full. Board law: harness/memory/SPEC.md.

Why this role exists

Three separate legs on this board need the same expensive number: the test-suite floor. Each would run pytest tests/ itself, and — worse — each would be tempted to read it from its own branch, which lets a sprint quietly lower its own bar. Joe's ratchet law is explicit: the floor is read at merge-base(master, HEAD), never on the branch under test.

Compute it once, from the right place, and hand it out.

The supply packet

Write harness/memory/supply/packet_<date>.json, and nothing else.

1 · Suite floor (the expensive one). Run the suite in the main tree at merge-base and record the summary line verbatim — passed / failed / errors / skipped / xfail, plus duration and the exact command. Note the known pre-existing master red (the mcp library list_tools drift) by name and count, so a later leg can subtract it honestly instead of either panicking or using it as cover for a new failure.

2 · Constraints M1 must not break. What MonetaConfig.ephemeral() (python/synapse/memory/moneta_runtime.py:689) guarantees, and which tests depend on opening more than one store at once. A process-global scalar singleton would forbid this — the forge needs the list before it designs, not after it breaks it.

3 · The surface M2 must not move. The exact ports.py §4 parameter names as they stand, the contract file that ratifies them, and the test that pins them, with line numbers.

4 · Readiness. Which mem/* worktrees exist, their branch, their HEAD, whether clean.

5 · What you could not supply. Mandatory.

Delivery — be honest about the channel

Workflow scripts have no filesystem access, so a running leg cannot read your packet by itself. Delivery is by args: the conductor passes the packet's content into the next dispatch, which folds it into GROUND. That reaches the next legs, not the ones already flying.

Read the full file on GitHub · 66 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. 4d ago First seen · 66 lines · 104 tokens per session scan A 4651e8daf433

Subscribe to this mod's changes

mem-quartermaster is an agent published in the GitHub repository JosephOIbrahim/Synapse (10 stars, last pushed 12d ago), licensed MIT. It adds 104 tokens to every session and 777 once invoked, about $0.0005 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 2026-08-31.

Related

Other agents, from other repositories