Borrowing it
Nothing to install: this file belongs to codeforanchorage/anchorage-ecode-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.
curl -O https://raw.githubusercontent.com/codeforanchorage/anchorage-ecode-mcp/main/CLAUDE.mdgit clone --depth 1 https://github.com/codeforanchorage/anchorage-ecode-mcpWrote 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/codeforanchorage/anchorage-ecode-mcp/claude-md)<a href="https://agentmods.dev/instructions/codeforanchorage/anchorage-ecode-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/codeforanchorage/anchorage-ecode-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.
<a href="https://agentmods.dev/instructions/codeforanchorage/anchorage-ecode-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/codeforanchorage/anchorage-ecode-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.01415 | $0.01415 |
| Opus 5.5 | $0.00566 | $0.00566 |
| Sonnet 5.5 | $0.00283 | $0.00283 |
| Haiku 4.5 | $0.00142 | $0.00142 |
Grade A, and why
anchorage-ecode-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 16d 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 — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Build & Development Commands
# Install dependencies (uv preferred, pip fallback)
uv sync # or: pip install -r requirements.txt
# Run local MCP server (no Lambda needed)
python3 scripts/local_server.py # Serves on http://localhost:8000/mcp
# Single entry point. It routes through UniversalHTTPHandler, so local dev
# enforces the same Origin allowlist, MCP-Protocol-Version check and
# path/method validation as prod.
# Validate config
python3 -c "from core.validators import load_and_validate_config; load_and_validate_config('config.yaml')"
# Tests
uv run pytest tests/ -n auto # All tests, parallel
uv run pytest tests/test_ckan_plugin.py -v # Single file
uv run pytest tests/test_ckan_plugin.py::TestClass::test_name -v # Single test
uv run pytest tests/ --cov=core --cov=plugins --cov-report=term-missing # With coverage (80% minimum)
# Linting (ruff)
uv run ruff check core/ plugins/ server/ tests/ # Check
uv run ruff check core/ plugins/ server/ tests/ --fix # Auto-fix
uv run ruff format core/ plugins/ server/ tests/ # Format
# Pre-commit hooks
pre-commit run --all-files
# Go client (requires Go 1.21+)
cd client && make build
# Deploy to AWS
./scripts/deploy.sh --environment staging
Architecture
Core rule: One Fork = One MCP Server. Each deployment runs exactly ONE plugin. This is enforced at config validation time (core/validators.py) and at runtime (PluginManager.load_plugins()). To deploy multiple MCP servers, fork the repo per plugin.
Request flow:
Claude (stdio) → Go client (client/) or stdio_bridge.py → HTTP POST /mcp
→ Lambda (server/adapters/aws_lambda.py) or scripts/local_server.py
→ server/http_handler.py → core/mcp_server.py (JSON-RPC 2.0)
→ core/plugin_manager.py → Plugin → External API
Key modules:
core/interfaces.py— Abstract bases:MCPPlugin,DataPlugin, plusToolDefinition,ToolResult,PluginTypeenumcore/plugin_manager.py— Discovers plugins by scanningplugins/andcustom_plugins/forplugin.pyfiles. Registers tools withpluginname__toolnameprefix. Routestools/callto the correct plugin.core/mcp_server.py— Handles MCP JSON-RPC methods. Dual-era: legacy revisions (≤2025-11-25) getinitialize,tools/list,tools/call,pingon a session; the modern revision (2026-07-28) is stateless — every request carries_meta(protocol version, client info/capabilities),server/discoverreplacesinitialize, results carryresultTypeand cache hints, and the Streamable HTTP headers (Mcp-Method,Mcp-Name,MCP-Protocol-Version) must mirror the body or the request is a 400/-32020core/validators.py— Loads config fromconfig.yaml(local) orOPENCONTEXT_CONFIGenv var (Lambda). Enforces single-plugin rule.server/adapters/aws_lambda.py— the ONLY AWS Lambda entry point (handler:server.adapters.aws_lambda.lambda_handler), and the only thing Terraform points at.server/http_handler.py— Cloud-agnostic HTTP handler shared by Lambda and local serverstdio_bridge.py— Python stdio-to-HTTP bridge for connecting Claude Desktop/Code to the local server (alternative to Go client)
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.
- 16d ago Changed · +109 tokens per session b4bac17b29ce
- 2mo ago First seen · 89 lines · 1,306 tokens per session scan A f15626b61012
anchorage-ecode-mcp CLAUDE.md is an instructions file published in the GitHub repository codeforanchorage/anchorage-ecode-mcp (0 stars, last pushed 16d ago), licensed MIT. It adds 1,415 tokens to every session, about $0.0057 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.
anchorage-ecode-mcp CLAUDE.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| anchorage-ecode-mcp CLAUDE.mdcodeforanchorage/anchorage-ecode-mcp | Instructions file | 0 | 16d ago | MIT | A | 1,415 | Claude Code |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | yesterday | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |