tracepass-mcp-server: Instructions file for Claude Code

CLAUDE.md

tracepass-mcp-server CLAUDE.md is an instructions file for Claude Code from malinoto/tracepass-mcp-server. It costs 3,016 tokens per session, scanned A, original, MIT.

A set of instructions for developing the tracepass-mcp-server, a public package that lets AI assistants manage products, EU Digital Product Passports, business partners, and GS1 EPCIS 2.0 supply-chain events.

In plain words
What is it for?
Use it when changing the server architecture, tool and result conventions, MCP features, or its separate release processes.
Why use it?
It keeps the server's tools and results consistent across its local package and hosted service, while protecting public code from private deployment details.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md.

This is malinoto/tracepass-mcp-server's own configuration. It tells Claude Code how to work on tracepass-mcp-server 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 tracepass-mcp-server configures →

Reuse

Borrowing it

Nothing to install: this file belongs to malinoto/tracepass-mcp-server. 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/malinoto/tracepass-mcp-server/master/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/malinoto/tracepass-mcp-server

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 tracepass-mcp-server CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/malinoto/tracepass-mcp-server/claude-md/github.svg)](https://agentmods.dev/instructions/malinoto/tracepass-mcp-server/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/malinoto/tracepass-mcp-server/claude-md"><img src="https://agentmods.dev/badge/instructions/malinoto/tracepass-mcp-server/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 tracepass-mcp-server CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/malinoto/tracepass-mcp-server/claude-md"><img src="https://agentmods.dev/badge/instructions/malinoto/tracepass-mcp-server/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 3,016 This file is loaded in full into every session.
When invoked 3,016 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 $0.03016 $0.03016
Opus 5.5 $0.01206 $0.01206
Sonnet 5.5 $0.00603 $0.00603
Haiku 4.5 $0.00302 $0.00302

Measured yesterday against content hash d184a46b764e, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

tracepass-mcp-server 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 yesterday.

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 · 180 lines

How it starts

The opening of the file, as written. The whole thing — 180 lines — stays where its author put it; the contents beside it link to each section on GitHub.

tracepass-mcp-server

MCP server for TracePass — lets AI assistants manage products, Digital Product Passports, economic-operator parties, and GS1 EPCIS 2.0 events over the TracePass v1 REST API. Speaks the full MCP protocol: tools, resources, resource templates, prompts. Read README.md for usage; this file is the agent guardrail layer.

This is a PUBLIC npm package (tracepass-mcp-server, integrators npx it). Nothing internal — no prod paths, SSH aliases, secrets, hosting details, or operational fingerprint — belongs in this repo. Deploy/ops artefacts and release keys live in private repos kept outside this one. Keep it that way when editing this file.

Architecture (the one rule everything serves)

One server core, every transport — the tool surface must never drift between the hosted endpoint and the local package.

  • src/server.ts — createMcpServer(config): builds the McpServer with all tools/resources/prompts. Transport-agnostic — knows nothing about stdio vs HTTP.
  • src/http.ts — hosted entrypoint (the Node service behind https://ai.tracepass.eu/mcp). Stateless: fresh server+transport per request, credential from Authorization: Bearer ….
  • src/stdio.ts — local/npx entrypoint. Client launches it as a subprocess; key from TRACEPASS_API_KEY env.
  • Both call the same createMcpServer. Only the transport + key source differ. Any change to the tool surface must go through the core so both forms stay identical.
  • src/usage-log.ts — one mcp_usage JSON line per hosted request on stdout: JSON-RPC method, tool + action, credential KIND (none / api_key / oauth / supplier_token), status, user-agent. Never the IP, the token or tool arguments. It is how real tools/call use is told apart from catalog and liveness probes.
  • src/supplier-server.ts — a second, separate server for suppliers answering one data request, served only by http.ts at /supplier/mcp/<token> (token in the path, for URL-only connectors) and /supplier/mcp (Bearer header). Its credential is the request token, not an account: no auth gate and no WWW-Authenticate challenge on that path, because an OAuth prompt is a dead end for someone with no TracePass account. A bad or missing token becomes a readable tool error (supplierResult). Its tools (the array returned in src/supplier-server.ts; the README's "For suppliers" table lists them — keep the two in lockstep) sit over the platform's /api/supplier/v1/*, which owns every rule (requested keys only, merge until review, rate limit). It is not in the eu.tracepass/tracepass listing, the server card or the main tool list, and its README table deliberately avoids the `name` — description pattern so mcppedia does not count its tools as the main server's.
  • src/compliance-server.ts — a third server, the DPP Assistant, served by http.ts at /compliance/mcp. It answers what EU DPP law asks of a company (scope, required fields, what changed, registry guidance, the saved compliance profile) and can start a draft passport; the tool array is buildComplianceTools(). Unlike the supplier path it authenticates like the main endpoint (tp_ key or OAuth, with its own WWW-Authenticate challenge and per-path RFC 9728 metadata at /.well-known/oauth-protected-resource/compliance/mcp, whose resource is exactly https://ai.tracepass.eu/compliance/mcp), and it has its own card at /.well-known/mcp/compliance-server-card.json, not mirrored on www. Every platform call carries X-Source: assistant, so answers are logged and capped on that channel. The platform owns every answer; the server adds two things only:
    • the hard-rule block (HARD_RULES) in each tool description and the server instructions: answer from a tool result, never from memory, never rank vendors;
    • the closing disclaimer: withClosingDisclaimer repeats the envelope's disclaimer as a separate final text block ("End your answer with this line, word for word"). Left only inside the JSON, models drop it from most answers. A tool argument must exist in the v1 request. /api/v1/scope/check is strict (an undeclared field or fact is 400 invalid_input), so an argument the tool advertises but the API lacks fails on every call that uses it. The model then retries without it, so nothing looks broken. A test checks scope_check's posted body against the published OpenAPI schema. Its registry entry is server.compliance.json (eu.tracepass/compliance), a separate listing from server.json. Its behaviour is measured by an eval set in the private runner repo; a tool-description change can move the scores, so re-run it after one.

Read the full file on GitHub · 180 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. yesterday Changed · -1 tokens per session d184a46b764e
  2. 5d ago Changed · +1 lines · +9 tokens per session 44c97f14b490
  3. 8d ago Changed · -1 lines · -5 tokens per session 88bab4d219a4
  4. 14d ago Changed · +2 lines · +40 tokens per session 402627bdd626
  5. 25d ago Changed · +27 lines · +508 tokens per session 40496678e56c
  6. 1mo ago Changed · +6 lines · +125 tokens per session fd33a581e4dc
  7. 1mo ago Changed · +4 lines · +89 tokens per session 6cc87319e3e1
  8. 1mo ago Changed · +12 lines · +222 tokens per session 2bf1657ec427
  9. 2mo ago First seen · 129 lines · 2,029 tokens per session scan A 3333015c73fd

Subscribe to this mod's changes

tracepass-mcp-server CLAUDE.md is an instructions file published in the GitHub repository malinoto/tracepass-mcp-server (1 stars, last pushed yesterday), licensed MIT. It adds 3,016 tokens to every session, about $0.0121 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

tracepass-mcp-server CLAUDE.md compared with instructions from other repositories

tracepass-mcp-server CLAUDE.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
tracepass-mcp-server CLAUDE.mdmalinoto/tracepass-mcp-serverInstructions file1yesterdayMITA3,016Claude Code
vibetags com-example-service-PricingService.instructions.mdPIsberg/vibetagsInstructions file25yesterdayMITA296GitHub Copilot
vscode buildNext.instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA9,719GitHub Copilot
next.js AGENTS.mdvercel/next.jsInstructions file143kyesterdayMITA7,461Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchainInstructions file149kyesterdayMITA4,680Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA5,516GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA3,093GitHub Copilot