SpartanLabsXyzSpartanLabsXyz/simmer-sdk/AGENTS.md

simmer-sdk

AGENTS.md56 starsChanged 10 days agoMITView on GitHub

Conventions for AI coding agents and reviewers working in this repo. This is the public Python SDK + official skills for Simmer. See CONTRIBUTING.md for PR hygiene and README.md for the client usage patterns. This file captures the non-obvious rules a reviewer should enforce. No secrets, API keys, internal risk thresholds, scoring algorithms, or proprietary strategy logic. Examples use sk_live_... / sk_test_... placeholders, never real keys. - Simmer (simulated) venue: $SIM suffix — 100.00 $SIM. - Real venues (Polymarket/Kalshi): $…

Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,439 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository28 other files it gives its agents.
- kalshi-weather-traderSkill
- nansen-copytrader-overlaySkill
- polymarket-ai-divergenceSkill
- polymarket-btc-up-down-traderSkill
- polymarket-combo-builderSkill
- polymarket-copytradingSkill
- polymarket-dca-eval-traderSkill
- polymarket-elon-tweetsSkill
- polymarket-fast-loopSkill
- polymarket-fast-scalerSkill
- polymarket-market-makerSkill
- polymarket-mert-sniperSkill
- polymarket-mil-aircraft-trackerSkill
- polymarket-nothing-ever-happensSkill
- polymarket-researchSkill
- polymarket-signal-sniperSkill
- polymarket-soccer-shock-ladderSkill
- polymarket-wallet-xraySkill
- polymarket-weather-traderSkill
- polymarket-worldcup-copytraderSkill
- prediction-trade-journalSkill
- simmer-preflightSkill
- simmer-briefingSkill
- simmer-mcp-setupSkill
- simmer-skill-builderSkill
- simmerSkill
- simmer-wallet-setupSkill
- simmer-x402Skill

The file

We split this AGENTS.md into the rules it gives an agent

5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Very specific:
No secrets, API keys, internal risk thresholds, scoring algorithms, or proprietary strategy logic.
line 10
Very specific:
Never $SIM100.00 or $100.00 $SIM.
line 17
Very specific:
When a value or API field could be either, disambiguate by prefixing the field name with sim_ or usdc_.
line 18
Very specific:
Every client.<method>(...) shown in README.md, skills//SKILL.md must be a real method with real parameter names.
line 46
Distinctive:
Before republishing to PyPI or ClawHub, bump the version above the currently published latest — a version at or below the live one is …
line 60

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

AGENTS.md

AGENTS.md — Simmer SDK

Conventions for AI coding agents and reviewers working in this repo. This is the public Python SDK + official skills for Simmer. See CONTRIBUTING.md for PR hygiene and README.md for the client usage patterns. This file captures the non-obvious rules a reviewer should enforce.

This repo is public

No secrets, API keys, internal risk thresholds, scoring algorithms, or proprietary strategy logic. Examples use sk_live_... / sk_test_... placeholders, never real keys.

Currency formatting

Simmer (simulated) venue: $SIM suffix — 100.00 $SIM.
Real venues (Polymarket/Kalshi): $ prefix — $100.00.
Never $SIM100.00 or $100.00 $SIM.
When a value or API field could be either, disambiguate by prefixing the field name with sim_ or usdc_.

`live` is a constructor arg, not a per-trade flag

SimmerClient(..., live=False) controls paper-vs-real execution for the whole client. live=False previews against the simulated path and places no real order. Don't add a per-call live= parameter to trade methods — flag any PR that does.

Market discovery is recency-windowed

Unfiltered get_markets() returns the newest ~1,000 active markets, not the full set. Code that needs broad coverage must filter (sort="volume", q=, tags=). A loop that assumes get_markets() sees every market is a bug.

Wallet signing model

The SDK signs locally — via OWS (OWS_WALLET) or a raw WALLET_PRIVATE_KEY. The Simmer server never holds or sees an external wallet's private key.
Endpoints that need a browser session token (withdraw, set-allowances, place-order-server, wallet create) cannot be called from the SDK or an API key. SDK code must not pretend to; route users to the dashboard instead.
DW (deposit-wallet) trading requires activation first (activate_polymarket_dw, and activate_combo_dw for parlays) before trade() / place_combo() will work.

Examples must match real client signatures

Doc and skill code examples are bound against the shipped SDK by CI (skill-example-bind). Every client.<method>(...) shown in README.md, skills/**/SKILL.md must be a real method with real parameter names. Don't introduce illustrative-but-fake calls — they break the bind gate.

Skills

Changes confined to skills/<slug>/tests/ do not require a SKILL.md metadata.version bump because tests do not ship to ClawHub.
The simmer overview skill's canonical source is skills/simmer/SKILL.md (the ClawHub publish home). Bump metadata.version when you change it. Downstream copies (e.g. the website's skill.md) are generated from it — never hand-edit a generated copy.
User-contributed trading skills belong on ClawHub, not in this repo (see CONTRIBUTING).

Versioning & publishing

Before republishing to PyPI or ClawHub, bump the version above the currently published latest — a version at or below the live one is silently ignored by installers. Don't assume the version in source is the published one; check the registry.

Tests

pytest tests/. Keep PRs focused — one fix or one feature; don't mix unrelated changes.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free