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 skills/myths-labs/muse/agent-protocolnpx skills add myths-labs/muse --skill agent-protocolgit clone --depth 1 https://github.com/myths-labs/museWrote 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/skills/myths-labs/muse/agent-protocol)<a href="https://agentmods.dev/skills/myths-labs/muse/agent-protocol"><img src="https://agentmods.dev/badge/skills/myths-labs/muse/agent-protocol.svg" alt="Measured on agentmods" 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 | $0.00025 | $0.02017 |
| Opus 5 | $0.00013 | $0.01009 |
| Sonnet 5 | $0.00005 | $0.00403 |
| Haiku 4.5 | $0.00003 | $0.00202 |
Grade A, and why
agent-protocol 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 5d 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 — 214 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Protocol Specification
Inspired by MemOS (multi-agent isolation + sharing) and multi-agent-memory (cross-agent persistence). Adapted for MUSE's pure Markdown zero-dependency architecture. Added in v2.12.0 (Batch 2 — competitive tech absorption).
Why
MUSE role files (.muse/*.md) currently work because AI Agents read Markdown and follow instructions. But as the ecosystem evolves toward multi-agent workflows (MCP servers, Claude Projects, multi-tool chains), role files need a machine-readable protocol — not just human-readable documents.
Problems this solves:
- Agent handoff: When switching between Agents (e.g., Claude → Gemini), context should be portable
- Multi-agent coordination: Multiple Agents working on the same project should respect role boundaries
- Tool interoperability: MCP servers, IDE plugins, and CLI tools should be able to read/write role files
Protocol: MUSE Role File Spec v1.0
File Discovery
.muse/
├── strategy.md # Strategy role
├── build.md # Build/Dev role
├── growth.md # Growth/Marketing role
├── qa.md # QA/Verification role
├── gm.md # General Manager role
├── ops.md # Operations role
├── research.md # Research role
├── fundraise.md # Fundraise role
└── archive/ # Archived content (not loaded)
Discovery rule: Any .md file in .muse/ (excluding archive/) = a role file.
L0 Header (Machine-Readable Status Line)
Every role file MUST have an L0 comment as its first line:
<!-- L0: {version} | {top_priority} | {status_flags} -->
Parsed fields:
| Field | Type | Description | Example |
|---|---|---|---|
version |
semver | Current project version | v2.11.0 |
top_priority |
string | Most important current task | P0=竞品技术吸收Batch2 |
status_flags |
csv | Comma-separated status indicators | QA PASS, 3 pending directives |
L0 is the Agent's "at a glance" API — any tool can grep all L0 lines to get project status in ~400 tokens.
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.
- 5d ago First seen · 214 lines · 25 tokens per session scan A 63f3905a5ec2
agent-protocol is a skill published in the GitHub repository myths-labs/muse (32 stars, last pushed 1mo ago), licensed MIT. It adds 25 tokens to every session and 2,017 once invoked, about $0.0001 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.
Other skills, from other repositories
project-brain
Triggers (all require explicit user request — do NOT activate just because a brain/ folder exists): (1) The user wants to set up project-level memory ("set up project brain", "scaffold project context", "init project brain", "建项目脑"). (2) The user is in a directory containing brain/ AND explicitly asks to resume /…
mex-inbox
Prepare and manage governed MEX Inbox proposals for durable Spec-family knowledge. Use when the user clearly asks to capture, create, save, or draft a MEX product Spec, requirement, constraint, acceptance criterion, or durable product-spec decision for team review, or explicitly invokes /mex-inbox or $mex-inbox. Do…
mex-relay
Prepare and manage durable MEX team handoffs. Use when the user asks to hand work to a person or team, prepare an end-of-session handoff, save what the next engineer needs, create a MEX Relay, take or close a Relay, or explicitly invokes /mex-relay or $mex-relay. Treat Relay as a memory and context baton, not chat…
context-receipts
Emit privacy-safe receipts for context selection, deferral, hydration, compaction, pruning, delegation, usage attribution, and boundary handoffs.
engram
Local-first personal AI identity and memory layer for MCP-compatible coding tools (Claude Code, Codex, Cursor, and others). Use this skill when the user wants to continue from a previous session ("continue from last session", "pick up where we left off"), recall a past decision ("what did we decide", "what was our…
evidence-attestation
Emit privacy-safe evidence attestations for agent actions, reviews, memory answers, handoffs, scores, approvals, and tool dispatches.