mcp-context-forge AGENTS.md

A set of instructions for coding agents working on IBM MCP ContextForge, an open-source system for connecting and governing tools, agents, and APIs. It explains the project structure, setup, commands, and area-specific guidance.

In plain words
What is it for?
Orienting an agent in the ContextForge codebase, finding the right project guidance, and following its development and validation procedures.
Why use it?
It gives an agent the repository context needed to make changes consistently. It also points to more specific instructions for tests, plugins, documentation, servers, and other parts of the project.

Instructions file for CodexOpenCode

Install

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.

agentmods
npx agentmods add instructions/ibm/mcp-context-forge/agents-md
Clone the repo
git clone --depth 1 https://github.com/IBM/mcp-context-forge

Made for: Codex, OpenCode.

Per session 7,788 This file is loaded in full into every session.
When invoked 7,788 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
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 $0.07788 $0.07788
Opus 5 $0.03894 $0.03894
Sonnet 5 $0.01558 $0.01558
Haiku 4.5 $0.00779 $0.00779

Measured 2d ago against content hash e3eff78e7da3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp-context-forge AGENTS.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.

Origin

Copies of this mod

2 near-identical copies found in the catalogue:

AGENTS.md · 576 lines

How it starts

The opening of the file, as written. The whole thing — 576 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Guidelines for AI coding assistants working with this repository.

For domain-specific guidance, see subdirectory AGENTS.md files:

  • tests/AGENTS.md - Testing conventions and workflows
  • plugins/AGENTS.md - Plugin framework and development
  • charts/AGENTS.md - Helm chart operations
  • docs/AGENTS.md - Documentation authoring
  • mcp-servers/AGENTS.md - MCP server implementation
  • crates/mcp_runtime/DEVELOPING.md - Rust MCP runtime development workflows, command matrix, and validation

Note: The llms/ directory contains guidance for LLMs using ContextForge solution (end-user runtime guidance), not for code agents working on this codebase.

Project Overview

ContextForge is an open source registry and proxy that federates MCP, A2A, and REST/gRPC APIs with centralized governance, discovery, and observability. It federates tools, agents, and APIs, optimizes agent and tool calling, and supports plugins, auth/RBAC, rate-limiting, virtual servers, multi-transport protocols, and an optional Admin UI.

Project Structure

mcpgateway/                 # Core FastAPI application
├── main.py                 # Application entry point
├── config.py               # Environment configuration
├── db.py                   # SQLAlchemy ORM models and session management
├── schemas.py              # Pydantic validation schemas
├── services/               # Business logic layer (50+ services)
├── routers/                # HTTP endpoint definitions (19 routers)
├── middleware/             # Cross-cutting concerns (16 middleware)
├── transports/             # Protocol implementations (SSE, WebSocket, stdio, streamable HTTP)
├── plugins/                # Plugin integration (uses cpex external package)
└── alembic/                # Database migrations

tests/                      # Test suite (see tests/AGENTS.md)
plugins/                    # Plugin implementations (see plugins/AGENTS.md)
plugins_rust/               # Rust plugin implementations for performance-sensitive paths
charts/                     # Helm charts (see charts/AGENTS.md)
docs/                       # Architecture and usage documentation (see docs/AGENTS.md)
a2a-agents/                 # A2A agent implementations (used for testing/examples)
mcp-servers/                # MCP server templates (see mcp-servers/AGENTS.md)
crates/                     # Direct Rust crate folders (runtime and wrapper)
llms/                       # End-user LLM guidance (not for code agents)

Read the full file on GitHub · 576 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. 2d ago First seen · 576 lines · 7,788 tokens per session scan A e3eff78e7da3

Subscribe to this mod's changes

mcp-context-forge AGENTS.md is an instructions file published in the GitHub repository IBM/mcp-context-forge (4,399 stars, last pushed today), licensed Apache-2.0. It adds 7,788 tokens to every session, about $0.0389 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 2026-08-30.