blindpay-mcp: Instructions file for Claude Code

CLAUDE.md

blindpay-mcp CLAUDE.md is an instructions file for Claude Code from blindpaylabs/blindpay-mcp. It costs 3,381 tokens per session, scanned A, original, MIT.

A maintenance guide for BlindPay MCP, a tool connection for the BlindPay public API. It describes commands and the automated process that regenerates the project when the API changes.

In plain words
What is it for?
Type-checking and building the project, synchronizing generated API code, checking generated changes, and validating tool-name limits.
Why use it?
It makes generated API updates repeatable and checks that the resulting code, package, and tool names remain valid.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: mentions .claude-plugin; mentions Claude Code; mentions Codex.

This is blindpaylabs/blindpay-mcp's own configuration. It tells Claude Code how to work on blindpay-mcp itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything blindpay-mcp configures →

Reuse

Borrowing it

Nothing to install: this file belongs to blindpaylabs/blindpay-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/blindpaylabs/blindpay-mcp/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/blindpaylabs/blindpay-mcp

Made for: Claude Code.

Wrote 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.

agentmods badge for blindpay-mcp CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/blindpaylabs/blindpay-mcp/claude-md/github.svg)](https://agentmods.dev/instructions/blindpaylabs/blindpay-mcp/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/blindpaylabs/blindpay-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/blindpaylabs/blindpay-mcp/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for blindpay-mcp CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/blindpaylabs/blindpay-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/blindpaylabs/blindpay-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 3,381 This file is loaded in full into every session.
When invoked 3,381 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.03381 $0.03381
Opus 5.5 $0.01352 $0.01352
Sonnet 5.5 $0.00676 $0.00676
Haiku 4.5 $0.00338 $0.00338

Measured 4d ago against content hash f3d5ca915fa5, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

blindpay-mcp 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 4d 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.

CLAUDE.md · 247 lines

How it starts

The opening of the file, as written. The whole thing — 247 lines — stays where its author put it; the contents beside it link to each section on GitHub.

BlindPay MCP - Agent Reference

Commands

npm run typecheck        # tsc --noEmit
npm run build            # tsc + chmod the entry point
npm run generate         # Raw openapi-mcp-generator dump to /tmp/mcp-generated (debugging only)
npm run sync              # Full deterministic pipeline: writes src/index.ts + package.json
npm run sync:check        # Same pipeline, but diffs against committed files instead of writing (exit 2 on diff)
npm run check-tool-names  # Static gate: fails if any committed tool name exceeds 52 chars
npm run check-plugin      # Static gate: fails if plugins/blindpay breaks a Claude directory rule
npm run check-openai-plugin  # Static gate: fails if plugins/openai breaks an OpenAI submission limit or leaves /mcp/readonly
npm run check-connector   # Live check: walks the OAuth sign-in of Claude, Claude Code, VS Code and ChatGPT against mcp.blindpay.com/mcp/readonly (not in CI)
npm test                  # Unit tests for scripts/lib (node:test via tsx)
npm run pack:openai       # ZIP for the ChatGPT/Codex plugin directory, in build-plugins/

How updates work

This repo is auto-synced with the BlindPay public API surface, the same set of routes shown on the public docs site (/doc). When that surface changes upstream, .github/workflows/api-sync.yml fires. There is no LLM anywhere in this path; every step is a plain script.

  1. The blindpay-v2 mcp-sync.yml workflow generates apps/api/openapi.json (already filtered by applyPublicTagFilter to public-tagged routes) and pushes it to this repo's api-sync-data branch at .api-sync/openapi.json. Then it fires a repository_dispatch api-sync event.
  2. api-sync.yml consumes the event and runs npm run sync (scripts/sync.mts), which:
    • runs openapi-mcp-generator against the spec,
    • extracts its toolDefinitionMap block,
    • deterministically shortens any tool name over 52 chars (see below),
    • rebuilds every tool's inputSchema from the spec the way the hosted server does (see "Parity with the hosted server"),
    • splices that map, plus the spec's webhook events, into src/index.template.ts to produce src/index.ts,
    • picks the version bump (see "Version bump") and writes that version into both package.json and SERVER_VERSION. Any spec shape the script can't map (missing map markers, a name it can't shorten under the limit, a duplicate name, a tool or parameter the generator and the hosted server's schema don't agree on) is a hard failure, not a best guess.
  3. CI (ci.yml, required on main) typechecks, builds and re-checks the 52-char limit. The PR on the api-sync branch is opened/updated and gh pr merge --auto --squash is set on it, so it merges itself once CI is green, no human step.
  4. publish.yml (unchanged) publishes to npm and tags a release on every push to main whose package.json version isn't already published.

Read the full file on GitHub · 247 lines

Changes

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.

  1. 4d ago Changed · +63 lines · +950 tokens per session f3d5ca915fa5
  2. 7d ago Changed · +14 lines · +230 tokens per session 8df6318255ae
  3. 11d ago Changed · +43 lines · +597 tokens per session fbcda7a3860c
  4. 1mo ago First seen · 127 lines · 1,604 tokens per session scan A 3e7f46b6b391

Subscribe to this mod's changes

blindpay-mcp CLAUDE.md is an instructions file published in the GitHub repository blindpaylabs/blindpay-mcp (9 stars, last pushed 4d ago), licensed MIT. It adds 3,381 tokens to every session, about $0.0135 per session on Opus 5.5. 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.

Related

Other instructions, from other repositories