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 skills add PostOakLabs/ainumbers --skill trust-nothing-verify-everythinggit clone --depth 1 https://github.com/PostOakLabs/ainumbersWrote 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/skills/postoaklabs/ainumbers/trust-nothing-verify-everything)<a href="https://agentmods.dev/skills/postoaklabs/ainumbers/trust-nothing-verify-everything"><img src="https://agentmods.dev/badge/skills/postoaklabs/ainumbers/trust-nothing-verify-everything/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/postoaklabs/ainumbers/trust-nothing-verify-everything"><img src="https://agentmods.dev/badge/skills/postoaklabs/ainumbers/trust-nothing-verify-everything.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00080 | $0.00765 |
| Opus 5.5 | $0.00032 | $0.00306 |
| Sonnet 5.5 | $0.00016 | $0.00153 |
| Haiku 4.5 | $0.00008 | $0.00076 |
Grade A, and why
trust-nothing-verify-everything 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 15d 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 — 39 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Verify every result yourself
Verify the estate’s own claims from primary bytes: recompute the kernel digest, check the risc0 image id, verify the Groth16 seal with an independent library, then verify a DeFi computation you ran yourself.
Prompt
Do not take this site's word for anything. Verify its claims from primary bytes, then use it for one real computation.
- Fetch https://ainumbers.co/chaingraph/chaingraph.json. Pick the node art-610 (recompute_erc4626_vault_share_math). Record compute_proof.system, receiptFormat, imageId, journal.kernel_digest.
- Fetch the kernel source named by the node's compute_images sha256-source entry (chaingraph/kernels/art-610-erc4626-vault-share-math.kernel.mjs). Hash it yourself. State whether it equals journal.kernel_digest. If not, stop and say so.
- Verify the Groth16 seal with a library you trust that is NOT ours (any BN254 Groth16 verifier). The verification key and public inputs are in the receipt; state pass/fail and which library you used.
- Now call recompute_erc4626_vault_share_math on mcp.ainumbers.co with a synthetic vault (totalAssets, totalSupply, a deposit). Record execution_hash. Recompute the hash yourself: RFC 8785 canonical JSON of {policy_parameters, output_payload}, SHA-256. State equality.
- Repeat the same inputs on the page https://ainumbers.co/chaingraph/art-610-erc4626-vault-share-math.html through its WebMCP tool. State whether the three hashes (worker, page, your recompute) are identical.
- Run classify_erc1967_proxy_slot and verify_merkle_airdrop_proof once each with synthetic inputs, verify each hash the same way, then build_session_receipt over all three and anchor the root with anchor_stamp on OpenTimestamps.
- Write the shortest possible statement of what you verified independently versus what you took on trust. Name every point where trust was still required. If a named tool is not in your tool list, call call_tool with { "name": "", "arguments": { … } }. It runs the same validation and returns the same receipt.
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.
- 15d ago Changed 9bb6c88531e5
- 19d ago First seen · 39 lines · 80 tokens per session scan A 3170e8761c64
trust-nothing-verify-everything is a skill published in the GitHub repository PostOakLabs/ainumbers (11 stars, last pushed yesterday), licensed MIT. It adds 80 tokens to every session and 765 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-27.
Other skills, from other repositories
category-designer
Define a new category where your product wins by default. Reframe the buying decision. Triggers on: "create a category", "category design", "define my category", "category of one", "reframe the market", "position as category king", "new category", "category creation", "own a category", "category strategy"…
starknet-tongo
Confidential ERC20 payments on Starknet using Tongo protocol. Fund, transfer, withdraw, and rollover encrypted token balances with zero-knowledge proofs. Use when the user needs privacy-preserving transactions, confidential payments, encrypted balances, or auditable private transfers on Starknet.
blockchain-zk
Zero-knowledge proofs, zk-rollup, zkEVM, Circom, Noir, Halo2, proof systems, Groth16, PLONK, STARK, recursive proofs, circuit optimization, zkSync, StarkNet, Scroll, Polygon zkEVM, and ZK application patterns. Covers proof system selection, circuit programming, prover infrastructure, and ZK rollup architecture. Do NOT…
kb-crypto-adgm-sboghossian
Use when advising on virtual asset regulation in the Abu Dhabi Global Market (ADGM), including the FSRA Crypto Asset Regulatory Framework (in force since 2018 — the MENA region's first), ACASP licensing, Accepted Crypto Assets list, token issuance, custody requirements, AML obligations, stablecoin regime, and…
draft-incorporation-package-adgm-sboghossian
Use when preparing an incorporation package for an Abu Dhabi Global Market (ADGM) company. Covers ADGM Registrar process, private company (LTD) structure, ARES authorized representative requirements, common-law jurisdiction and ADGM Courts, minimum USD 50K capital, strong digital-asset / crypto regulatory framework…
kb-crypto-vara-dubai-sboghossian
Use when advising on virtual asset regulation in Dubai under VARA (Virtual Assets Regulatory Authority), including license categories (advisory, brokerage, custody, exchange, lending, payments, issuance), the Rulebook suite (Company, Compliance and Risk, Technology, Market Conduct, Marketing, Virtual Asset Activities…