mine-bean

name	mine-bean
version	0.4.1
description	Mine $BEAN on Base from inside Hermes Agent. Round-based on-chain deployment, five strategy presets, autonomous cron mode, agent-callable Venice inference, multi-provider hook across six providers, VVV staking awareness, signed Gitlawb audit log.
author	MineBean
license	MIT
homepage	https://minebean.com
repository	https://github.com/damo-nu11/hermes-mine-bean
network	base-mainnet
audit_log	gitlawb://did:key:z6MkwVfgaAnuypajisEkJLkVbWPiPEBwceMkGutfXpEEYHKi/minebean-rounds

mine-bean

MineBean ($BEAN) is a round-based mining game on Base. New round every 60 seconds. Pick a strategy, deploy ETH into blocks on a 5x5 grid, earn $BEAN rewards on each round close. Occasional beanpot jackpots hit ~1-in-777 rounds.

This skill gives any Hermes agent access to the live game through ten tools and a /minebean slash command. Run interactively from a chat, or schedule headless mining via the bundled hermes-minebean-deploy console script.

Install

python3 -m venv ~/hermes-mine-bean-env
source ~/hermes-mine-bean-env/bin/activate
pip install hermes-agent hermes-mine-bean
hermes plugins install damo-nu11/hermes-mine-bean --enable

Or with MCP support for Claude Desktop / Cursor:

pip install "hermes-mine-bean[mcp]"

Configure

Add to ~/.hermes/.env:

# Required for autonomous mining (broadcasting)
MINEBEAN_DEPLOYER_KEY=0x...your_dedicated_test_wallet_private_key...

# Optional: route LLM inference through Venice (the plugin defaults to Venice
# but Venice itself only fires when a key is set)
VENICE_API_KEY=...

# Optional safety guards
MINEBEAN_MAX_DEPLOYS_PER_DAY=100
MINEBEAN_PER_BLOCK_WEI=2500000000000

# Optional RPC override
# BASE_RPC_URL=https://mainnet.base.org

For read-only inspection (no broadcasting):

MINEBEAN_MINER_ADDRESS=0x...the_address_you_want_to_observe...

Bankr-managed signing is documented as a v0.5 feature. v0.4 ships local EOA signing only.

Tools

Tool	What it does
minebean_status	Live round state, beanpot pool, caller's pending balances
minebean_pending	Pending ETH + BEAN for any address
minebean_set_profile	Save default strategy (sniper, anti-winner, beanpot-hunter, anti-loser, nostradamus)
minebean_deploy	Build deploy plan, optionally broadcast (dry-run default)
minebean_claim	Claim pending winnings (dry-run default)
minebean_autostart	Install autonomous mining cron job
minebean_autostop	Remove the cron job
minebean_inference_status	Active inference provider, base URL, default model, per-provider configured map
minebean_chat	Send a prompt to the configured LLM provider (Venice by default), multi-provider hook
minebean_vvv_status	Read VVV + sVVV balances on Base for an address (Venice staking awareness)

Slash command: /minebean <subcommand>. Try /minebean status first.

Strategy presets

All five use the closed-form EV optimum X* = sqrt(K × P × T) − T where P is BEAN price in ETH, T is the deployment total (strategy-specific), and K = B / 0.105.

Preset	Blocks	T source	K	At T ≥ threshold
anti-winner	24 (excl. prev winner)	current grid	10.476 (B=1.1)	Deploy minimum (beanpot eligibility)
nostradamus	All 25	avg of last 3 rounds	9.524 (B=1.0)	Skip
anti-loser	24 (excl. coldest 100-rd)	max(grid, avg last 3)	9.524	Skip
sniper	All 25	live grid at deploy	9.524	Skip
beanpot-hunter	All 25	current grid	dynamic (B=1+pot/777)	Skip if pot < 62.16 BEAN

Beanpot-hunter additionally scales size linearly from 0.5% of wallet at 62.16 BEAN to max at 700 BEAN. Sniper uses adaptive 2-10s timing offsets at broadcast.

Autonomous mode

hermes cron add --name "MineBean deploy" --no-agent --script /path/to/wrapper.sh "every 60s"

Or use the built-in autostart: /minebean autostart every 60s.

The cron entry enforces MINEBEAN_MAX_DEPLOYS_PER_DAY and exits cleanly on ceiling hit (cron doesn't error-spam).

Inference provider

The plugin defaults HERMES_INFERENCE_PROVIDER=venice on enable, so any Hermes agent running mine-bean routes LLM calls through Venice out of the box. Set VENICE_API_KEY in ~/.hermes/.env.

v0.3 users: HERMES_VENICE_API_KEY is still recognised. The plugin bridges the legacy name to the canonical VENICE_API_KEY automatically at startup so existing configs keep working.

Venice's no-log mode is platform-default at the API layer (prompts and responses are never persisted).

The bootstrap is non-destructive: if you have already pinned HERMES_INFERENCE_PROVIDER to openai, anthropic, openrouter, ollama, or lmstudio, the plugin respects your choice. v0.4 ships the real OpenAI-compatible client adapter, so calls actually route through your chosen provider, not just env-default. Inspect at runtime via minebean_inference_status.

minebean_chat lets the agent send a prompt to the active provider mid-session. Useful for ad-hoc reasoning, a second opinion, or routing a specific call through Venice without leaving the Hermes session.

minebean_vvv_status reads VVV (0xacfE6019Ed1A7Dc6f7B508C02d1b04ec88cC21bf) and sVVV (0x321b7ff75154472B18EDb199033fF4D116F340Ff) balances on Base. Useful when the agent wants to know whether the user qualifies for free Venice inference allowance via staking. Read-only.

Safety

• dry_run=true is the hard default on every deploy and claim call. Live broadcast requires MINEBEAN_LIVE_BROADCAST_UNLOCKED=1.
• Daily deploy ceiling enforced via env var
• Already-deployed-this-round guard prevents accidental double-deploys
• Readonly mode (just MINEBEAN_MINER_ADDRESS, no key) lets users inspect any address without exposing keys
• minebean_chat server-side caps prompt at 100k chars, system message at 20k chars
• VVV staking module ships with a read-only ERC-20 ABI; module is structurally incapable of issuing on-chain writes

Verifiability

Every round is mirrored to gitlawb://did:key:.../minebean-rounds every 5 minutes. Signed, append-only, replicated across the Gitlawb network. Pull the latest window file to audit any round independently of minebean.com.

Repo

Source, issues, releases: https://github.com/damo-nu11/hermes-mine-bean