NOTES

An experimental protocol extension that lets a server advertise a list of specialist agents to another program. The program can discover an agent's description and capabilities before requesting its detailed instructions and tools.

In plain words
What is it for?
Use it to discover available agents, inspect one agent's details, and route tool calls to the appropriate specialist.
Why use it?
It avoids loading every specialist agent and all of their tool definitions at once. That makes agent discovery more selective for systems that coordinate many agents.

Agent

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 agents/panyam/mcpkit/notes
Clone the repo
git clone --depth 1 https://github.com/panyam/mcpkit
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,049 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00000 $0.01049
Opus 5 $0.00000 $0.00524
Sonnet 5 $0.00000 $0.00210
Haiku 4.5 $0.00000 $0.00105

Measured yesterday against content hash 07cd52cdce88, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

NOTES 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.

experimental/ext/agents/NOTES.md · 97 lines

How it starts

The opening of the file, as written. The whole thing — 97 lines — stays where its author put it; the contents beside it link to each section on GitHub.

experimental/ext/agents — implementation notes

Server-declared agent discovery (agents-wg issue 20, pre-SEP). For the API see README.md.

Research surface under experimental/ext/, with its own go.mod, mirroring experimental/ext/events. Promote to ext/agents only when a SEP merges. Like experimental/ext/events, it is not in the per-PR test.yml matrix. It runs via the experimental umbrella and make testall only.


What it is

A server hosting a fleet of specialist agents advertises them as a small roster of tuples for routing, so a supervisor host never eager-loads a flat tools/list of every specialist's schemas.

Three-level progressive disclosure, the same shape as two-tier skills:

  1. capabilities.extensions["io.mcpkit/agents"], advertised via the core.ExtensionProvider mechanism
  2. agents/list → a roster of AgentSummary (agentId, description, capabilities, exampleTasks, delegateTool, tasksEnabled, skillUri), with no tool schemas
  3. agents/get {agentId}AgentDetail (the summary embedded, plus instructions and scoped tools[])

Only discovery is new wire surface. Invocation rides the existing tools/call via each agent's advertised delegateTool.


Why the identifier is io.mcpkit/agents, not io.modelcontextprotocol/agents

It used to be the latter. The value tracked the Agents WG's working name, which read as harmless while the primitive was clearly experimental.

It is not harmless, because the WG has not chosen among the three wire shapes its research doc lists (RPC, resources under agent://, or an extension) and mcpkit ships the RPC shape. Matching the name while the shape is unsettled is the worse of the two failure modes: a client that recognises the standard identifier negotiates it and then receives a payload that does not match what the eventual spec defines. A client that does not recognise io.mcpkit/agents simply skips the primitive, which is correct behaviour for an unratified surface. Not matching fails safely; matching wrongly does not.

Read the full file on GitHub · 97 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 First seen · 97 lines · 0 tokens per session scan A 07cd52cdce88

Subscribe to this mod's changes

NOTES is an agent published in the GitHub repository panyam/mcpkit (5 stars, last pushed 2d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 1,049 tokens. 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 agents, from other repositories

Demonstrate

Agent for demonstrating VS Code features.

microsoft/vscode · 10 tokens

playwright-test-generator

Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.

microsoft/playwright · 151 tokens

.NET-Notebook-Migration-Agent

Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.

microsoft/ai-agents-for-beginners · 33 tokens

AVM Owner Triage

Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.

github/awesome-copilot · 61 tokens

Ultimate Transparent Thinking Beast Mode

Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.

github/awesome-copilot · 11 tokens

code-reviewer

Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.

anthropics/claude-cookbooks · 52 tokens