poker-strategy

name	poker-strategy
description	PokerBot internal dependency — GTO poker calculation tools and strategy docs. This is a pure internal component that should NEVER be triggered directly by users. Only coachbot and bot-management SKILLs call this SKILL's tools (equity, odds, preflop, evaluator) and strategy docs during analysis/decisions. If the user asks "EV多少" or "胜率", that triggers coachbot, not this SKILL.
author	EnyanDai
version	2.0.0
tags	["poker","gto","strategy","tools","internal"]
metadata	{"openclaw":{"requires":{"bins":["python"]},"install":[{"kind":"pip","package":"numpy"}]}}

Poker Strategy — Router

Tools

Use the app-pinned Python from paths.env for every tool call:

source ./paths.env 2>/dev/null || true
"${PY:-python}" .agents/skills/poker-strategy/tools/<tool>.py ...

Do not call bare python3; it may resolve to the system Python, which does not have PokerBot's app dependencies such as numpy.

Tool	Use when...	Example command	Returns
preflop	Open / 3-bet / call decision for a hand + position	`source ./paths.env 2>/dev/null
equity	% win vs a villain range (optional board)	`source ./paths.env 2>/dev/null
odds	Is this call +EV given pot / call / equity?	`source ./paths.env 2>/dev/null
evaluator	Final hand rank from 5–7 cards	`source ./paths.env 2>/dev/null

Each tool also accepts --help for full arg list.

Strategy docs

Read on-demand from .agents/skills/poker-strategy/strategy/<name>.md:

Doc	Read when wondering...
preflop	"Is KQo a 3-bet vs BTN?" / "How wide do I defend BB vs CO?"
postflop	"Who has range advantage on K72r?" / "C-bet this flop or check back?"
sizing	"Why 1/3 pot here and not 2/3?" / "What bluff freq does half-pot allow?"
gto-fundamentals	"What's the MDF facing 2/3 pot?" / "When should I deviate from balance?"
range	"What combos is villain continuing to barrel with?" / "Has their range capped?"

Card notation

Ranks 2-9 T J Q K A · Suits h d c s · e.g. Ah=A♥ Td=T♦