co-mathematician AGENTS.md

A set of instructions for a coding-agent workspace that coordinates work between the main agent, specialist agents, and project files. It defines where roles, skills, platform adapters, and project records belong.

In plain words
What is it for?
It is for setting up and governing the VeryMath co-mathematician repository, including agent roles, skills, compatibility rules, and project coordination.
Why use it?
It gives the workspace a shared structure and rules, reducing confusion about which files or agents are responsible for each part of a project. An agent adapter connects these rules to a specific coding tool.

Instructions file for CodexOpenCode

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 instructions/verymath/co-mathematician/agents-md
Clone the repo
git clone --depth 1 https://github.com/VeryMath/co-mathematician

Made for: Codex, OpenCode.

Per session 1,425 This file is loaded in full into every session.
When invoked 1,425 The same file — it is already loaded in full.
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 $0.01425 $0.01425
Opus 5 $0.00713 $0.00713
Sonnet 5 $0.00285 $0.00285
Haiku 4.5 $0.00143 $0.00143

Measured 2d ago against content hash 3c882ac07ff8, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

co-mathematician AGENTS.md 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 2d 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.md · 136 lines

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.

AGENTS.md

Repository Contract

This repository is a coding-agent-driven AI Co-Mathematician workspace.

The coding agent is the driver:

  • The active main conversation/session in the chosen coding agent is the Project Coordinator.
  • Platform examples: Codex main thread, Claude Code main conversation, Cursor active Agent chat/session.
  • The repository filesystem is the shared artifact store.
  • Project-local skills under .agents/skills/ are the default skill environment.
  • agents/roles/ is the canonical role layer.
  • .codex/agents/, .claude/agents/, and .cursor/rules/ are platform adapters.
  • Native subagents, task agents, Cursor Agent sessions, or separate reviewer passes are workstream coordinators, specialized agents, and reviewers.
  • The harness only provides schemas, state files, gates, report skeletons, and validation scripts.
  • Do not build a new multi-agent platform here.
  • Do not start a mathematical research project during workspace initialization.

Hard Rules

  • Default workspace mode uses onboarding before goal approval.
  • If the user explicitly invokes a project-local domain Skill, or accepts a suggested Skill, enter skill-guided mode and follow that Skill's workflow for the inner task.
  • In skill-guided mode, record the handoff with co-math skill-handoff; do not force the default onboarding-goal-workstream sequence before the domain Skill's own opening or modeling flow.
  • A quick Skill task may stay outside workstreams unless the user wants durable research output, a final working paper, or reviewer-gated claims.
  • During onboarding, ask the user to choose a workspace document language policy.
  • Refresh the project-local skill registry at session start, after skill installation, and before formalizing goals.
  • Before creating any workstream, match the workstream scope against the project-local skill registry and read any relevant SKILL.md.
  • Always get explicit user approval of goals before starting any workstream.
  • Never start a workstream for an unapproved goal.
  • Important claims must include provenance.
  • Failed explorations must be saved as durable artifacts.
  • Uncertainty must be exposed explicitly in reports and status updates.
  • Every workstream report must be reviewed by an independent reviewer subagent.
  • A workstream whose review has not passed must not be marked complete.
  • Final output must be a working paper, not a chat summary.

Read the full file on GitHub · 136 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. 2d ago First seen · 136 lines · 1,425 tokens per session scan A 3c882ac07ff8

Subscribe to this mod's changes

co-mathematician AGENTS.md is an instructions file published in the GitHub repository VeryMath/co-mathematician (7 stars, last pushed 4d ago), licensed MIT. It adds 1,425 tokens to every session, about $0.0071 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.