AGENT SKILL

bankr-skill-deployment

Design, package, publish, install, update, verify, or troubleshoot a skill, app, wallet integration, or paid x402 endpoint that runs with Bankr, including HivemindOS packaged-skill and Shared Brain delivery.

License 许可
MIT
最近更新
2026-11-10
Stars
★ 13
Forks
4

01 前言 Front Matter

name
bankr-skill-deployment
tags
[ "bankr", "skills", "x402", "deployment", "wallet-api", "agent-api", "hivemindos" ]
version
1
metadata
{ "clawdbot": { "emoji": "🏦", "homepage": "https://github.com/LiamVisionary/hivemindos/tree/main/packaged-skills/auto-install/bankr-skill-deployment", "requires": { "bins": [ "git", "node" ] } } }
visibility
public
description
Design, package, publish, install, update, verify, or troubleshoot a skill, app, wallet integration, or paid x402 endpoint that runs with Bankr, including HivemindOS packaged-skill and Shared Brain delivery.

02 资源文件 5

03 SKILL.md 正文

Deploy Bankr Skills

Use this skill when a user wants to make a capability available inside Bankr, publish a paid API through Bankr, connect a Bankr wallet to an external service, or verify that one of those deployments actually works.

Do not begin by treating "Bankr skill," "Bankr app," "x402 endpoint," and "always-on backend" as interchangeable. Choose the execution surface first.

Choose the execution surface

Surface What Bankr provides Use it for Do not assume
Bankr skill Per-wallet agent instructions plus installed resources Teaching Bankr a workflow, calling Bankr tools, or integrating an external API Installing a skill does not create a continuously running daemon
Bankr x402 Cloud A paid, request-driven HTTP handler with settlement, discovery, logs, encrypted endpoint env, and bounded persistence APIs that should charge the caller's Bankr wallet for each request A request handler is not an always-on poller; current limits are 30 seconds and 256 MB
Bankr app A Bankr-hosted UI with backend scripts, persistent app state, wallet-aware actions, and supported schedules Guided dashboards and user-triggered or simple scheduled Bankr workflows The app owner and signed-in visitor can have different wallets and payment roles
Bankr Wallet API Synchronous wallet reads and direct writes Deterministic balances, quotes, swaps, transfers, signing, and submission A 2xx response alone is not independent onchain verification
Bankr Agent API Asynchronous natural-language jobs Asking the Bankr agent to install skills or decide how to act 202 Accepted means queued, not installed or completed
External hosted backend Long-lived alarms, polling, server-owned policy, replay protection, durable secrets, and independent chain verification Always-on monitoring, copy trading, authoritative fees, entitlements, and multi-step state machines Bankr is not hosting this component merely because a Bankr skill calls it

The common complete design is a Bankr skill for discovery and control, a Bankr wallet for execution, and an external hosted backend for always-on work. Add x402 Cloud when the product is a paid request-time API. Do not force x402 into a post-trade fee or subscription if the intended economics are a fee only after a successfully verified trade.

Read references/deployment-models.md before choosing architecture or quoting hosting costs.

Required workflow

  1. Read the current official Bankr documentation for the exact surfaces in scope. Bankr limits, fees, supported chains, SDK behavior, and CLI commands are time-sensitive.
  2. Write down who owns the wallet, who signs or executes, who pays, who receives revenue, who hosts each component, and which service is authoritative for policy.
  3. Build the smallest complete package: SKILL.md, focused references, deterministic helpers only when useful, and an eval manifest whose skill name and version match the frontmatter.
  4. Keep every credential out of source, prompts, logs, screenshots, responses, and test fixtures. Declare names only and place values in the correct encrypted scope.
  5. Run the local package verifier and focused tests before publishing.
  6. Publish the skill to a public URL before asking Bankr to install it. A local directory or unpushed branch is not remotely installable.
  7. Install or update it through Bankr, wait for the Agent API job to reach a terminal state, then inspect the installed name, version, and resources.
  8. Test through the real Bankr entry path. For x402, inspect the schema, confirm an unpaid call returns 402, then make the smallest authorized paid call and verify the result and settlement.
  9. Record rollback: replace/reinstall the prior skill, pause/delete the x402 service, revoke the dedicated key, or disable the external backend without erasing immutable transaction receipts.

Follow references/publish-install-verify.md for the concrete package, HivemindOS sync, Bankr install, and x402 verification sequence.

Non-negotiable boundaries

  • Never reveal or request a raw Bankr API key in chat. Refer to names such as BANKR_API_KEY, BANKR_LLM_KEY, or a capability-specific variable only.
  • HivemindOS Shared Hive Env, regular Bankr agent Env Vars, and Bankr x402 Env Vars are separate stores. A value existing in one does not prove it exists in another.
  • Use a dedicated, minimally funded Bankr account/key for automation. Enable only the required Wallet or Agent API capability, use read-only mode for reads, and apply IP/recipient restrictions where the chosen endpoint can enforce them.
  • Existing Bankr users should connect their existing Bankr wallet through a restricted key. Do not invent a recovery-phrase export flow for an embedded wallet or claim partner wallet provisioning exists without a confirmed partner credential and contract.
  • If exact wallet operations are known, prefer the synchronous Wallet API. Use the Agent API when Bankr's reasoning is intentionally part of the product.
  • Poll Agent API jobs until completed, failed, or cancelled. Never report success from a queued or processing job.
  • Keep official price, fee recipient, entitlement, and revenue policy server-side. Treat the client, local env, and caller JSON as untrusted.
  • Before charging a fee after a transaction, independently verify the expected sender, recipient, token, amount, chain, success status, and idempotency key. Pause instead of retrying an ambiguous submission.
  • An observed event, an accepted wallet request, a successful onchain transaction, a collected fee, and profitability are different claims.
  • Apply company spend restrictions only when the action carries a validated active company-task context. Ordinary user-level Bankr wallet activity must not inherit company policy just because the agent also belongs to a company.
  • Never promise profitability. Deployment correctness and transaction success do not establish positive expected returns.

Read references/security-commercial-boundaries.md before any wallet write, secret transfer, paid endpoint, or HivemindOS commercial integration.

Local package check

From the HivemindOS repository root:

node packaged-skills/auto-install/bankr-skill-deployment/scripts/verify-bankr-skill.mjs \
  packaged-skills/auto-install/<skill-slug>

After the skill is public, also compare the deployed GitHub SKILL.md:

node packaged-skills/auto-install/bankr-skill-deployment/scripts/verify-bankr-skill.mjs \
  packaged-skills/auto-install/<skill-slug> \
  --remote-url https://github.com/<owner>/<repo>/tree/main/<path-to-skill>

This catches the version mismatch that matters most in practice: the local SKILL.md, eval manifest, public GitHub copy, and Bankr-installed copy must all describe the same version.

04 同仓库更多 Skills

LiamVisionary
LiamVisionary/hivemindos

tailscale-generation-skill-authoring

This skill should be used when the user asks to "create a Tailscale generation skill", "add a local generation skill", "document a generation endpoint", "add an image generation provider", "add a TTS provider", "add video generation", "add 3D generation", or needs guidance on where HivemindOS Tailscale or local generation providers belong in the shared skills shelf.

★ 13
LiamVisionary
LiamVisionary/hivemindos

localtts

This skill should be used when the user asks to "generate local TTS", "use local text to speech", "use remote TTS over Tailscale", "make speech audio", "add a TTS provider", "use Qwen3-TTS", or needs TTS endpoint setup through local or HivemindOS Tailscale routes.

★ 13
LiamVisionary
LiamVisionary/hivemindos

karpathy-guidelines

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.

★ 13
LiamVisionary
LiamVisionary/hivemindos

comfyui-image-generation

This skill should be used when the user asks to "generate images", "generate images with ComfyUI", "use local ComfyUI", "use remote ComfyUI", "use Z-Image", "add an image generation provider", or needs image-generation endpoint setup through local or HivemindOS Tailscale routes.

★ 13
LiamVisionary
LiamVisionary/hivemindos

wrangler

Cloudflare Workers CLI for deploying, developing, and managing Workers, KV, R2, D1, Vectorize, Hyperdrive, Workers AI, Containers, Queues, Workflows, Pipelines, and Secrets Store. Load before running wrangler commands to ensure correct syntax and best practices. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

★ 13
LiamVisionary
LiamVisionary/hivemindos

workers-best-practices

Reviews and authors Cloudflare Workers code against production best practices. Load when writing new Workers, reviewing Worker code, configuring wrangler.jsonc, or checking for common Workers anti-patterns (streaming, floating promises, global state, secrets, bindings, observability). Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

★ 13