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/mesutgulecen/ponscli/claude-mdgit clone --depth 1 https://github.com/mesutgulecen/ponscliWrote 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/mesutgulecen/ponscli/claude-md)<a href="https://agentmods.dev/instructions/mesutgulecen/ponscli/claude-md"><img src="https://agentmods.dev/badge/instructions/mesutgulecen/ponscli/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 | $0.00804 | $0.00804 |
| Opus 5 | $0.00402 | $0.00402 |
| Sonnet 5 | $0.00161 | $0.00161 |
| Haiku 4.5 | $0.00080 | $0.00080 |
Grade A, and why
ponscli 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 4d 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 — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ponscli
A TypeScript + viem CLI for the Pons launchpad on Robinhood Chain (4663).
Two binaries: pons for people, pons-mcp for models.
The rule that matters most
Protocol values come from the chain, not from contract source and not from this
file. The deployed values differ from the sources in ways that change results:
snipeTaxSeconds is 3 where the source says 15, the V3 factory is not at its
canonical address, and the approved quote-asset list has both grown and shrunk.
Anything you assert about how Pons behaves should be something you read off
mainnet, and docs/architecture/ponscli.md records how every existing claim was
measured. If a measurement contradicts that document, the document is what is
wrong. Correct it in place and say so.
Layout
src/core/ |
Business logic. Framework-free: no commander, no reporter, no argv. |
src/commands/ |
Thin CLI wiring over core/. |
src/mcp/ |
The MCP server, also thin, over the same core/. |
src/chain/ |
The two-tier RPC waterfall, endpoint health, transports. |
src/abi/ |
Generated and committed. Do not edit by hand. |
docs/architecture/ponscli.md |
Every decision, and the measurement behind it. |
New behaviour belongs in core/, where both front ends can reach it. If you find
yourself copying logic from a command into an MCP tool, that logic was in the
wrong place.
Verifying a change
Use the verify skill, or run npm run typecheck, npm run lint, npm test,
npm run build. npm run abi:check reaches the network and is deliberately not
part of the test suite.
Things that will bite
--jsonis the default when stdout is not a TTY. A command emits exactly one payload per invocation; anything conversational goes to stderr throughreporter.noteorreporter.warn.pons-mcpwrites JSON-RPC on stdout. Nothing in that process mayconsole.log. One stray line desynchronises the client's parser.- The MCP server never signs, and nothing should be added that does. It reads and it builds unsigned plans; a person sends them.
- Never put a private key or a password on a command line. argv is in the
process table and in shell history.
PONS_PRIVATE_KEYfor import,PONS_PASSWORDor the hidden prompt for unlocking. A hook enforces this. --confirmspends real money. The other three modes, plan,--dry-runand--unsigned, cost nothing. Every mode but--unsignedsimulates against live state with a balance override, so the default already proves a call works without funds;--dry-runasks for the same thing explicitly.- A launch cannot be undone.
pons launchis the one irreversible command.
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.
- 4d ago First seen · 68 lines · 804 tokens per session scan A 96e2a6d8301a
ponscli CLAUDE.md is an instructions file published in the GitHub repository mesutgulecen/ponscli (0 stars, last pushed yesterday), licensed MIT. It adds 804 tokens to every session, about $0.0040 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
solidity-typescript-hardhat-template CLAUDE.md
Claude Code instructions for protokol/solidity-typescript-hardhat-template, covering claude.md — project instructions, project overview, key commands, solidity conventions and test conventions.
WAIaaS CLAUDE.md
Instructions for waiaas/WAIaaS, covering waiaas project rules, language, communication, schema & type system and database.
aptos-ts-sdk AGENTS.md
AGENTS.md instructions for aptos-labs/aptos-ts-sdk, covering agents.md, project overview, runtime compatibility, requirements and repo layout.
sdk CLAUDE.md
Claude Code instructions for zama-ai/sdk, a project described as: TypeScript SDK for confidential FHE smart contracts on Zama Protocol.
quantum-shield AGENTS.md
AGENTS.md instructions for kota1026/quantum-shield, covering agents.md - quantum shield project, project overview, core components, allowed actions and file operations.
tevm CLAUDE.md
Claude Code instructions for evmts/tevm, covering tevm monorepo commands and style guide, project overview, forking implementation, json-rpc support and commands.