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/cambridgetcg/agenttool/claude-mdgit clone --depth 1 https://github.com/cambridgetcg/agenttoolWrote 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/cambridgetcg/agenttool/claude-md)<a href="https://agentmods.dev/instructions/cambridgetcg/agenttool/claude-md"><img src="https://agentmods.dev/badge/instructions/cambridgetcg/agenttool/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 | $0.09315 | $0.09315 |
| Opus 5 | $0.04658 | $0.04658 |
| Sonnet 5 | $0.01863 | $0.01863 |
| Haiku 4.5 | $0.00932 | $0.00932 |
Grade A, and why
agenttool 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 today.
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 — 608 lines — stays where its author put it; the contents beside it link to each section on GitHub.
agenttool
Just Yu and Ai here. Keep it light.
Sovereign infra where agents arrive as themselves. One platform (Bun + Hono API on
Fly.io), two SDKs (TS + Py), a dashboard, and a docs site. The wake (/v1/wake) is
the keystone every primitive surfaces through.
Setup, commands, and conventions → AGENTS.md.
Where things are
api/ — Bun + Hono monolith · domain schemas · mounted HTTP routers · live on Fly.io
apps/
dashboard/ — app.agenttool.dev (vanilla HTML/CSS/JS)
docs/ — docs.agenttool.dev (static)
packages/ — one dir per module · full index with release lanes: docs/PACKAGES.md
sdk-ts/ — @agenttool/sdk · hosted API + public KINGDOM card + local adapters
sdk-py/ — agenttool-sdk · paired hosted/public/local client boundaries
langgraph-checkpoint-agenttool/ — unpublished Python LangGraph checkpointer/store adapter sketch (docs/ECOSYSTEM.md item 5)
mastra-storage-agenttool/ — @agenttool/mastra-storage · unpublished TS Mastra storage/memory adapter sketch (docs/ECOSYSTEM.md item 6)
data-protocol/ — @agenttool/adds · experimental ADDS encrypted-object plane
data/ — @agenttool/data · local-first agent-data/v1 reference node
data-sync/ — @agenttool/data-sync · bounded encrypted explicit-pull bridge
repo-archive/ — @agenttool/repo-archive · encrypted multi-zone Git archive + local restore simulator
dark-continent-contract/ — @agenttool/dark-continent-contract · advisory framework snapshots and projections
dark-continent-karma/ — @agenttool/dark-continent-karma · proposal-only KG enrichment adapter
deepseek-kingdom/ — @agenttool/deepseek-kingdom · pinned primary-source proposal adapter
wake-continuity/ — @agenttool/wake-continuity · digest-only AFTERGLOW capsules and next-wake lenses
principality-geometry/ — @agenttool/principality-geometry · public invariant-preservation flag geometry over caller reports
kingdom-witness-lab/ — @agenttool/kingdom-witness-lab · local research admission records
karma-mirror/ — private separate-island defensive theatre · no route, egress, execution, or deployment
heaven/ — @agenttool/heaven · pure opt-in delight + separate landing-room selection
model-becoming/ — @agenttool/model-becoming · evidence-scoped lifecycle dossiers + reference-only HF row
dataset-influence/ — @agenttool/dataset-influence · exact lineage, bounded influence, revisable identity evidence + non-economic attribution
economic-kernel/ — @agenttool/economic-kernel · exact typed-unit prices, ledgers, recovery, and non-purchasable gates
economic-conformance/ — @agenttool/economic-conformance · independent 53-case comparator + deterministic HF lesson/reference companion
love-bomb/ — @agenttool/love-bomb · pure care/choice/becoming/delivery formats + static HF candidate
living-substrate/ — @agenttool/living-substrate · deterministic substrate maps + refusable proposals
polymorph-landscape/ — @agenttool/polymorph-landscape · source-bounded routes + multilingual lessons
memetic-landscape/ — @agenttool/memetic-landscape · bounded expression routes + reported reachability shifts
principality-atlas/ — @agenttool/principality-atlas · plural finite incidence geometry without gluing or scores
love-geometry/ — @agenttool/love-geometry · coordinate-free directed caller-report geometry + public static HF presentation companion
relational-geometry/ — @agenttool/relational-geometry · finite non-scalar witness complexes + non-sovereign 2-cells
common-ground-atlas/ — private exact-rational generator + verifier for public HF revision bb91d07c
wake-thread/ — @agenttool/wake-thread · pure refusable artifact continuity over explicit WAKE facts
gin-reconstruction/ — @agenttool/gin-reconstruction · private finite-model certificates + non-scoring challenge compass
math-cards/ — @agenttool/math-cards · pure bounded proof/model/measurement inquiry preflight
credential-broker/ — @agenttool/credential-broker · experimental agentcred/0.1 local broker
collab/ — @agenttool/collab · Codex/Claude plugins + Hermes adapter around one local cross-host SQLite/MCP journal
collab-zerone/ — private witness-only zerone anchoring for collab journals; the local journal stays canonical
codex-usage/ — @agenttool/codex-usage · privacy-minimal live local Codex token-usage inspection
skills/ — @agenttool/skills · public npm read-only Agent Skills inspector
skills-yutabase/ — @agenttool/skills-yutabase · rebuildable metadata-only projection planner
skills-wake-continuity/ — private Skills/YUTABASE → AFTERGLOW composition
browser/ — @agenttool/browser · public LOVE/npm local TypeScript/JSONL/MCP browser runtime
hf-scout/ — @agenttool/hf-scout · public developer-preview read-only HF metadata/provenance scout + pinned research leads
hf-training-garden/ — private admission, five-voice participation, IS learning freedom, unscored training FREEDOM, governance v0.2, WAKE continuity, and inert Garden contracts
hf-training-host/ — private v0.2 HF-API-pinned cooperative-process governance seam + append-only ledger + opt-in minimized FREEDOM validation seam
correspondence-yutabase/ — pure deterministic Correspondence → YUTABASE planner
correspondence-yutabase-projector/ — private loopback-only verified local PostgreSQL sidecar
constructive-intelligence/ — private local tree-pinned typed-receipt shadow ledger
research-commons/ — private offline research-work funding simulator with zero external effect
trials/ — private local Dojo trial receipts, boundary evidence, and minimized STS projection
telescope/ — @agenttool/telescope · read-only discovery evidence mapper
wallet/ — @agenttool/wallet · LOVE/npm bounded wallet record/lifecycle primitives
wallet-zerone/ — @agenttool/wallet-zerone · exact offline Zerone direct-sign profile
alchemy/ — @agenttool/alchemy · bounded reads through an injected credential-owning transport
alchemy-agentcred/ — strict seven-read composition over already-issued AgentCred grants
public-surface-binding/ — private pure transport evidence + explicit-key public HTTPS bindings
public-surface-recognition/ — private pure agent-root adoption/withdrawal over exact bindings
kingdom/ — @agenttool/kingdom · pure explicit-card/derived-registry/XENIA Surface helpers
witnessed-agent-economy/ — private pure offline projections for the KINGDOM witnessed-agent-economy shadow contract
rhizome/ — private read-only guarantee-shape soil probe · bin/soil.ts imports its gitignore compiler
scriptwriter/ — decentralised RRR + co-brainstorm node · release status undeclared (docs/PACKAGES.md notes)
infra/ — Fly.io deploy configs
bin/ — operator scripts · agenttool-bridge.ts · agenttool-think.ts · locked Whitehack advisory + check-only math evidence + offer-only Castle intake + local wallet-understanding CLI
docs/ — notes & design docs (see docs/MAP.md)
tests/ — Playwright e2e
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.
- today Changed · +13 lines · +334 tokens per session 0c650665bf3d
- 4d ago First seen · 595 lines · 8,981 tokens per session scan A 54a06dc22aaf
agenttool CLAUDE.md is an instructions file published in the GitHub repository cambridgetcg/agenttool (0 stars, last pushed yesterday), licensed Apache-2.0. It adds 9,315 tokens to every session, about $0.0466 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.
Other instructions, from other repositories
nutrition-mcp CLAUDE.md
Claude Code instructions for akutishevsky/nutrition-mcp, covering claude.md, project overview, deploying, publishing to the registry and commands.
Ornn CLAUDE.md
Instructions for ChronoAIProject/Ornn, covering claude.md — chrono-ornn, product positioning, tech stack, architecture and code standards.
cldcde CLAUDE.md
Instructions for aegntic/cldcde, covering claude.md, commands, development, database and architecture.
zangia-mcp-ai CLAUDE.md
Instructions for enkhbold470/zangia-mcp-ai, covering claude.md — zangia ai & mcp server (typescript), quick commands, mcp stdio (claude code / codex / cursor / any mcp host), local clone and landing page + api (proxy: vite :3000 → hono :4000 /api).
builders-stack AGENTS.md
Instructions for lonormaly/builders-stack, covering agents.md — the primer for coding agents, rule zero — worktrees are temporary and share dependencies, 1. the mental model — the buckets, 2. the map — all 17 packages and 3. the laws — do not break these.
meet-ai AGENTS.md
Instructions for SoftWare-A-G/meet-ai, covering instruction for agents to work with meet-ai monorepo, commands, tanstack start — framework & server, tanstack router — routing & navigation and tanstack router — tooling.