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 instructions/markmhendrickson/ateles/claude-mdgit clone --depth 1 https://github.com/markmhendrickson/atelesWrote 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/instructions/markmhendrickson/ateles/claude-md)<a href="https://agentmods.dev/instructions/markmhendrickson/ateles/claude-md"><img src="https://agentmods.dev/badge/instructions/markmhendrickson/ateles/claude-md.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 | $0.09393 | $0.09393 |
| Opus 5 | $0.04696 | $0.04696 |
| Sonnet 5 | $0.01879 | $0.01879 |
| Haiku 4.5 | $0.00939 | $0.00939 |
Grade A, and why
ateles CLAUDE.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.
How it starts
The opening of the file, as written. The whole thing — 224 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ateles — Claude Code Project Instructions
Plan and task maintenance (automatic)
Each session maintains the Neotoma plan entity that matches its own workstream — never a fixed, hardcoded plan. The swarm-architecture plan ent_99ace4dd6673aa36ed08b1fe ("Ateles Agent Swarm Architecture") is the plan for swarm-architecture work only. Unrelated workstreams (tax prep, Neotoma release engineering, website, cloud hosting, etc.) each have their own plan and MUST NOT write into the swarm plan. Writing one workstream's decisions/todos into another's plan is the collision that corrupted this plan in June 2026.
Select the bound plan once per session, as soon as the workstream is clear:
- Resolve the matching plan:
mcp__mcpsrv_neotoma__retrieve_entitieswithentity_type: planand asearchfor the workstream; pick the closest match. - If no existing plan fits, create one (
/update-planskill) and bind to it. - Maintain only that bound plan for the rest of the session.
Apply these rules on every turn, without being asked — to the bound plan:
- Before correcting
decisionsortodos, RE-READ the current field and MERGE.correctreplaces the entire field, so add or update only the keys/items you authored and preserve every entry already present. NEVER rebuild a field from a stale in-memory copy — that silently deletes other sessions' entries. If the field changed since you last read it, re-read and re-merge before writing. - When a todo item is completed — correct its
statusto"done"intodos, with relevant entity IDs, file paths, or PR numbers in anotesfield. - Never mark a task or todo
donewhile citing a commit, branch, file, or PR that does not resolve. Verify the artifact exists first (gitcat-file/ls-remote, GitHub). Unverifiable completion claims are the second failure mode that corrupted this plan in June 2026. - When a decision is settled — add or correct one entry in the
decisionsmap. One sentence, snake_case key. - When blockers change (something unblocked, something newly blocked) — correct
next_stepsto reflect the current state. - When a new actionable task is identified — create a
taskentity and link itPART_OFthe bound plan. Use theupdate-tasksskill for field values and priority mapping. - When a daemon, entity, or file is renamed — correct any stale references in
body,decisions, andtodosin the same turn the rename happens.
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.
- 2d ago First seen · 224 lines · 9,393 tokens per session scan A fb4bfd12ed20
ateles CLAUDE.md is an instructions file published in the GitHub repository markmhendrickson/ateles (6 stars, last pushed yesterday), licensed MIT. It adds 9,393 tokens to every session, about $0.0470 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 2026-09-03.
Other instructions, from other repositories
harness-sdk AGENTS.md
AGENTS.md instructions for strands-agents/harness-sdk, covering agent development guide - strands agents monorepo, context, monorepo layout, where the "why" lives: team/ and writing code.
matryca-plumber AGENTS.md
AGENTS.md instructions for MarcoPorcellato/matryca-plumber, covering agents.md — matryca plumber instruction router, audience map, instruction loading strategy, always-on working agreements and rule and skill routing.
matryca-plumber CLAUDE.md
Claude Code instructions for MarcoPorcellato/matryca-plumber, covering code audit — code intelligence, always do, never do, resources and cli.
tokenjam AGENTS.md
AGENTS.md instructions for Metabuilder-Labs/tokenjam, covering agents.md — tokenjam and quick rules for agents that don't read claude.md.
hermes-desktop CLAUDE.md
Claude Code instructions for fathah/hermes-desktop, covering before starting work, post-task checklist (required — do not skip), what is lat.md?, commands and syntax primer.
oneclaw CLAUDE.md
Instructions for 0xNyk/oneclaw, covering claude.md, build & development commands, architecture, core traits (the extension points) and adding a new implementation.