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/polymind-inc/agent-framework-js/claude-mdgit clone --depth 1 https://github.com/polymind-inc/agent-framework-jsWhat 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 | $0.00775 | $0.00775 |
| Opus 5 | $0.00387 | $0.00387 |
| Sonnet 5 | $0.00155 | $0.00155 |
| Haiku 4.5 | $0.00077 | $0.00077 |
Grade A, and why
agent-framework-js 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 — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Notes for AI coding assistants
This repository is a TypeScript implementation of the Microsoft Agent Framework programming model,
written so that it could be contributed upstream to microsoft/agent-framework. Read
CONTRIBUTING.md first — it is the source of truth for how changes are made here.
The points below are the ones most often gotten wrong.
Non-negotiable rules
- Semantics match the reference implementations exactly. Signatures follow TypeScript idiom; behaviour does not get to. Default values, side effects, folding, loop behaviour, cancellation and serialization must match. When implementations disagree, the order of authority is .NET, then Python, then Go. Read the reference source — never infer behaviour, and never guess.
- Use the current (v1.13) naming generation.
Agent,AgentSession,Message,Content,AgentResponse. The pre-rename names —ChatAgent,AgentThread,AgentRunResponse,ChatMessage— appear nowhere, including comments and tests. - Do not break the wire format. Serialized properties are camelCase;
Contentdiscriminator values are the Python snake_case literals; unrecognized data is preserved and round-tripped. - The core has exactly one runtime dependency,
@opentelemetry/api. Schema libraries are accepted through the Standard Schema interface, never depended on. No Node-specific APIs in the core — it has to keep running on Deno, Bun, edge runtimes and in browsers. - API surface changes are grounded in the reference implementations; new features are gated.
The public API is no longer frozen, but a change to its surface must be justified by alignment
with the reference implementations — name the reference behaviour it matches. Breaking changes
are called out in the changelog (during
0.x, minor releases may break). New functionality — a new package, a new API area, workflows — is not built without an explicit request or an issue asking for it; quality, performance and parity work take priority. - Fix bugs reproduction-first. Write a test that fails against the current code, make the fix, then temporarily revert the fix and confirm the test fails again.
- Comments are self-contained. They explain the code to a first-time reader. No references to internal design documents, milestone identifiers or decision logs.
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 First seen · 50 lines · 775 tokens per session scan A 4e47f58448bd
agent-framework-js CLAUDE.md is an instructions file published in the GitHub repository polymind-inc/agent-framework-js (4 stars, last pushed 2d ago), licensed MIT. It adds 775 tokens to every session, about $0.0039 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
adk-go AGENTS.md
AGENTS.md instructions for google/adk-go, covering agents.md, project overview, skills, setup & core commands and definition of done.
adk-go CLAUDE.md
Claude Code instructions for google/adk-go: See AGENTS.md for project context, commands, and contribution guidelines for AI coding agents.
LeAgent AGENTS.md
Instructions for vixues/LeAgent, covering leagent development guidelines, project overview, architecture, execution topology (one kernel, many ingresses) and surface map (where things live).
lucid-agents AGENTS.md
Instructions for daydreamsai/lucid-agents, covering lucid agents monorepo - ai coding guide, project overview, brand system, architecture overview and package dependencies.
semantix CLAUDE.md
Instructions for Gnosil/semantix, covering 项目规则（claude code） and 1. 改动须经用户过目；由 claude 提交到分支 + pr，用户 review 后合并.
agentarea AGENTS.md
Instructions for agentarea/agentarea, covering agents.md, where to look, structure, commands and backend (from agentarea-platform/).