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/echennells/sparkbtcbot/claude-mdgit clone --depth 1 https://github.com/echennells/sparkbtcbotWrote 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/echennells/sparkbtcbot/claude-md)<a href="https://agentmods.dev/instructions/echennells/sparkbtcbot/claude-md"><img src="https://agentmods.dev/badge/instructions/echennells/sparkbtcbot/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 5.1 | $0.01583 | $0.01583 |
| Opus 5 | $0.00792 | $0.00792 |
| Sonnet 5 | $0.00317 | $0.00317 |
| Haiku 4.5 | $0.00158 | $0.00158 |
Grade A, and why
sparkbtcbot 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 2d 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 — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
sparkbtcbot
Claude Code skill for setting up Spark Bitcoin L2 wallet capabilities for AI agents.
Installation: the plugin marketplace, or clone anywhere (e.g. ~/sparkbtcbot) and symlink the nested skill: ln -s ~/sparkbtcbot/skills/sparkbtcbot ~/.claude/skills/sparkbtcbot. Never clone directly into ~/.claude/skills/ — the skill nests at skills/sparkbtcbot/, so a direct clone buries SKILL.md too deep to be discovered (README → Local clone).
What This Skill Does
Teaches Claude Code how to give AI agents Bitcoin capabilities using the Spark L2:
- Initialize Wallet — Create or import a BIP39 mnemonic-based wallet
- Check Balance — Query BTC and token balances
- Receive Deposits — Generate L1 Bitcoin deposit addresses
- Transfer BTC — Instant, zero-fee Spark-to-Spark transfers
- Lightning Invoices — Create and pay BOLT11 invoices (Lightning Network interop)
- Spark Invoices — Native Spark invoices payable in sats or tokens
- Token Operations — Transfer BTKN/LRC20 tokens natively
- Withdraw to L1 — Cooperative exit back to on-chain Bitcoin
- Message Signing — Sign and verify messages for identity proof
- Merchant Purchases — Buy real-world goods/services with sats (Bitrefill, nadanada) under a shared payment policy (
references/merchant-spending.md): invoice-vs-quote guard, confirm-before-buy, PII consent, bearer-secret deliverables - Unilateral-Exit Backup — Continuously-fresh
spark.unilateral-exit-bundle.v1recovery bundle (the "leaf-vault"), consumed by Blink's spark-unilateral-exit tool if the operators ever vanish
Structure
lib/
atomic-file.js # the one atomic writer (temp+fsync+link/rename+dir-fsync)
encrypted-seed.js # scrypt + AES-256-GCM seed file helper
leaf-vault.js # SDK-free recovery-bundle persistence + shape validation
fee-guards.js # fee/amount ceilings for sends, claims, withdrawals
spend-ledger.js # rolling-window cumulative budget (bounds send LOOPS)
transfer-ids.js # persisted per-invoice Lightning dedup transferId (safe retries)
recipients-allowlist.js # opt-in outbound allowlist guardrail
index.js / index.d.ts # npm entry (also exports ./leaf-vault subpaths)
skills/
sparkbtcbot/
SKILL.md # Always-loaded skill body (security, setup, navigator)
references/ # Detail loaded on demand (SDK API, agent class, L402, etc.)
encrypted-seed.md # Threat model, setup modes, recovery
unilateral-exit.md # The leaf-vault backup + Blink's exit tool
recovery-scenarios.md # Recovery properties (staleness, justice, economics)
scripts/ # Runnable example scripts
cli.js # `sparkbtcbot <command>` — the single published bin (dispatcher)
setup-encrypted-seed.js # `npm run setup` — one-time bootstrap
leaf-vault.js # snapshotLeafVault / verifyVault / enableLeafVault (library)
leaf-vault-cli.js # `npm run leaf-vault [-- verify]` — snapshot/verify CLI
balance-and-deposits.js
payment-flow.js
token-operations.js
spark-agent.js
l402-paywalls.js
evals/ # Skill-quality evals (does the skill make Claude produce correct/safe code?)
evals.json # Output evals: SDK-correctness + security-behavior, with checkable assertions
trigger-eval.json # Description-triggering queries (see NOTES: not measurable via claude -p here)
NOTES.md # How to run (subagent output evals, with-skill vs baseline) + last results
tests/ # vitest suite for the LIBRARY code (unit, integration, funded tiers)
.env.example # Environment variable template
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.
- 2d ago Changed · +86 tokens per session 390efcb316d8
- 6d ago First seen · 96 lines · 1,497 tokens per session scan A b47a0a2d5680
sparkbtcbot CLAUDE.md is an instructions file published in the GitHub repository echennells/sparkbtcbot (4 stars, last pushed 3d ago), licensed MIT. It adds 1,583 tokens to every session, about $0.0079 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
ln-agent-poc-v2 CLAUDE.md
Claude Code instructions for cachet-jp/ln-agent-poc-v2, covering claude.md, project, layout, common commands and architecture notes.
SaturnZap copilot-instructions.md
Instructions for lqwdtech/SaturnZap, covering saturnzap — copilot instructions, build & test, architecture, conventions and test patterns.
SaturnZap tests.instructions.md
Use when writing, editing, or debugging SaturnZap tests. Covers fixtures, mocking patterns, CLI runner usage, and live test markers.
crypto-claude-skills AGENTS.md
Instructions for Vo1ganin/crypto-claude-skills, covering agents.md, skills, operating rules, read-only by default and credentials.
coingecko-mcp-server AGENTS.md
AGENTS.md instructions for cyanheads/coingecko-mcp-server, covering developer protocol, what's next?, core rules, patterns and tool.
coingecko-mcp-server CLAUDE.md
Claude Code instructions for cyanheads/coingecko-mcp-server, covering developer protocol, what's next?, core rules, patterns and tool.