Borrowing it
Nothing to install: this file belongs to Franlinozz/Sluice. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/Franlinozz/Sluice/main/CLAUDE.mdgit clone --depth 1 https://github.com/Franlinozz/SluiceWrote 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/franlinozz/sluice/claude-md)<a href="https://agentmods.dev/instructions/franlinozz/sluice/claude-md"><img src="https://agentmods.dev/badge/instructions/franlinozz/sluice/claude-md/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/instructions/franlinozz/sluice/claude-md"><img src="https://agentmods.dev/badge/instructions/franlinozz/sluice/claude-md.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 | $0.07178 | $0.07178 |
| Opus 5.5 | $0.02871 | $0.02871 |
| Sonnet 5.5 | $0.01436 | $0.01436 |
| Haiku 4.5 | $0.00718 | $0.00718 |
Grade A, and why
Sluice 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 2mo 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 — 344 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SLUICE — Project Rules & Constants (read this fully every session)
First action of every session: read this entire file. Everything you build obeys it. Working tree is
/root/Sluice(this is NOT Tessera/Xyndicate/Agora/Apogee/Archon/Auralis — never touch those projects). Repo:Franlinozz/Sluice. Vercel project:sluice(prj_P12m6QjGi3gFtPSCfXbDTiGJRbi9, teamfranlinozzs-projects).
What we're building
Sluice is a settlement layer for the agent-paid web. Any unit of value — a read, a second, a citation, a listen, an API call — is metered and settled on Arc in USDC. Creators get paid per use; agents pay per use and DECIDE for themselves. Hero use case: AI research agents pay creators PER CITATION (the open, creator-owned alternative to Cloudflare Pay-Per-Crawl / RSL, which declare terms but don't settle).
Non-negotiable rules
- NEVER fake data, balances, figures, transactions, or "working" integrations. A clearly labeled "beta"/"coming soon"/"roadmap" state ALWAYS beats a fake. Every number shown to a user must trace to a real on-chain (Arc testnet) event or a real DB record.
- NO DEAD CONTROLS. Every button/link/toggle either works or is visibly disabled with a stated reason (tooltip/label). The product owner is not a deep dev and cannot spot a silent no-op — so there must be none.
- Network-agnostic settlement. All chain/settlement access goes through ONE interface (packages/chain). Arc Testnet is primary; swapping the RPC provider (or, last resort, another network) must be a config change, never a code rewrite. Do NOT scatter RPC URLs/chain IDs.
- Meter decoupled from settlement backend. The accrual engine must not care whether the final settle is Gateway-batched or a direct x402 call. If Gateway-on-Arc misbehaves, we can fall back to direct x402 settlement without touching accrual logic.
- USDC DECIMALS DISCIPLINE (critical):
- ERC-20 USDC used for PAYMENTS = 6 decimals. All payment amounts, EIP-3009 values, and ledger math use 6-decimal integer base units. Use viem parseUnits(amount, 6).
- Arc's NATIVE GAS token is USDC displayed with 18 decimals. Never mix gas display with payment amounts. Keep a single money helper (packages/money) and route all amounts through it.
- EIP-3009 is signed against the GATEWAY WALLET CONTRACT, not the USDC token contract.
The x402
schemefield is"exact"; the string"GatewayWalletBatched"is the EIP-712 DOMAIN NAME placed inextra.name, withextra.verifyingContract= the Gateway Wallet. (This is the #1 thing teams get wrong — see "x402 payment requirements shape" below.) - One-time deposit required: a wallet must deposit USDC into the Gateway Wallet contract BEFORE it can make nanopayments. Build and surface this step explicitly.
- Settlement LAG is real: seller earnings appear only AFTER Circle Gateway settles the batch on-chain (can take minutes). UI must show honest states: authorized → batching → settled. Never show settled balance the instant a payment is authorized.
- Money in code is always integer base units (bigint). Format for display only at the edge. JSON cannot carry bigint — serialize bigints as strings, parse back at boundaries.
- Verify before assuming. When unsure of an API shape, address, or version, READ the canonical source (links below) or the arc-nanopayments repo — do not invent signatures.
- Commit after every green Definition of Done, with a clear message. Keep PR-sized diffs.
- Secrets: server-only keys live in env, never in client bundles. In Next.js, only NEXT_PUBLIC_* is exposed to the browser. Private keys/API keys must NEVER be NEXT_PUBLIC_.
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.
- 2mo ago First seen · 344 lines · 7,178 tokens per session scan A a98902d85488
Sluice CLAUDE.md is an instructions file published in the GitHub repository Franlinozz/Sluice (36 stars, last pushed 2mo ago), licensed MIT. It adds 7,178 tokens to every session, about $0.0287 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 2026-08-30.
Sluice CLAUDE.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| Sluice CLAUDE.mdFranlinozz/Sluice | Instructions file | 36 | 2mo ago | MIT | A | 7,178 | Claude Code |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | yesterday | MIT | A | 7,461 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 5,649 | GitHub Copilot |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 3,093 | GitHub Copilot |
| vscode ai-customization.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 1,907 | GitHub Copilot |