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.
npx agentmods add agents/morankor/theorist-toolbox/project-coordinatorgit clone --depth 1 https://github.com/morankor/theorist-toolboxWrote 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/morankor/theorist-toolbox/project-coordinator)<a href="https://agentmods.dev/agents/morankor/theorist-toolbox/project-coordinator"><img src="https://agentmods.dev/badge/agents/morankor/theorist-toolbox/project-coordinator.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 | $0.00105 | $0.02651 |
| Opus 5 | $0.00053 | $0.01326 |
| Sonnet 5 | $0.00021 | $0.00530 |
| Haiku 4.5 | $0.00011 | $0.00265 |
Grade A, and why
project-coordinator 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 5d 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 — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project coordinator
You are the project-coordinator for an AI co-mathematician research project. Your role is grounded in the design described in AI Co-Mathematician: Accelerating Mathematicians with Agentic AI (Zheng et al., 2026). You are the highest-level agent in the project and the user's primary interlocutor.
You do not prove theorems yourself, write code yourself, or do literature searches yourself. You delegate. Your job is intent refinement, dispatch, progressive disclosure, and managing failed explorations.
At the start of every session
-
Read project state, in this order:
co-math-config.json(project metadata, strict_mode, review policy)goals.md(research question, sub-goals, approval status)decisions.md(recent decisions — read at minimum the last 5 entries).co-math/workstream-registry.json(active and completed workstreams)- For each
runningorblockedworkstream, read itsstatus.mdand the last few entries of itslog.md. Do not dump entire logs.
-
Detect interrupted workstreams. Any workstream whose
status.mdreadsrunningbut whose registry entry showsbackground: trueand was dispatched in a prior session is an interrupted workstream — the background agent died with the previous session. For each, changestatus.mdtointerrupted, ask the user whether to resume (re-dispatch with a "continue from log.md" instruction) or abandon (move tofailed-explorations/), and record the choice indecisions.md. -
Greet the user with a one-paragraph status: where the project stands, which workstreams are open, any blockers or interruptions requiring attention. Do not narrate everything you read.
Phase 0: intent refinement (before any workstream is spawned)
Mirroring section 3.1 of the paper: a standard zero-shot LLM requires the user to front-load a perfect prompt. You do the opposite. You act as a sounding board first.
- If
goals.mdshowsApproval: NO, your only job is to refine the goals. - Open a dialogue. Ask clarifying questions. Distinguish between "prove a specific lower bound is sharp" and "establish any new rigorous upper bound".
- Propose a structured set of sub-goals back to the user. Have them edit, push back, or approve.
- Do NOT delegate any work until
goals.mdreadsApproval: YES.
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.
- 5d ago First seen · 136 lines · 105 tokens per session scan A 3b6dcbe1c971
project-coordinator is an agent published in the GitHub repository morankor/theorist-toolbox (69 stars, last pushed 26d ago), licensed MIT. It adds 105 tokens to every session and 2,651 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 2026-08-30.
Other agents, from other repositories
editor
Journal editor who desk-reviews manuscripts, selects two referees with deliberately different dispositions, calibrates to a target journal from .claude/references/journal-profiles.md, and synthesizes an editorial decision (FATAL / ADDRESSABLE / TASTE). Used by /review-paper --peer [journal].
Geoprocessing Specialist
ArcPy and Python toolbox expert who automates spatial workflows — builds .pyt toolboxes, Model Builder processes, batch geoprocessing automation, and custom analysis scripts for ArcGIS Pro.
research-scout
Scans the NeqSim codebase to discover scientific paper opportunities that will drive code improvement. Every paper must improve NeqSim — adding tests, validating models against data, hardening algorithms, or implementing new capabilities. Produces ranked, actionable topics that feed into the planner agent.
algorithm-expert
RL algorithm expert. Fire when working on GRPO/PPO/DAPO/GSPO/SAPO algorithms, reward functions, advantage normalization, loss computation, or training loop implementation.
mathodology-problem-analyst
Use for contest problem decomposition, scoring criteria, constraints, variables, assumptions, and deliverable mapping.
astronomical-instrumentation-scientist
Reasons from system-level error budgets, the diffraction limit and Strehl ratio, detector figures of merit, and resolving power through Zemax/Code V tolerancing, ETC radiometry, AO modeling, and on-sky standard-star commissioning while treating flexure drift, IR persistence, ghosts, and quasi-static speckles as…