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/hypercycle-development/mosaic-companion/agents-mdgit clone --depth 1 https://github.com/hypercycle-development/mosaic-companionWhat 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 | $0.02534 | $0.02534 |
| Opus 5 | $0.01267 | $0.01267 |
| Sonnet 5 | $0.00507 | $0.00507 |
| Haiku 4.5 | $0.00253 | $0.00253 |
Grade A, and why
mosaic-companion 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 3d 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 — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MosAIc Companion — AI Agent Instructions
This file provides context for AI coding assistants working on this project. Read this before making changes to the codebase.
Project Overview
MosAIc is an Electron desktop app (React + TypeScript) that serves as an AI companion with tool use, wallet integration, and secure sandboxed tool execution.
Stack: Electron (main process) + React (renderer) + TypeScript. Vite for bundling.
Architecture — MUST READ Before Modifying
The project follows a Core (trusted) vs Sandbox (untrusted) architecture for tool execution.
Full architecture docs: /docs/architecture/
Key Rules
- WASM is the primary runtime for sandboxed tools. Docker is optional for heavy workloads (GPU, databases). WASM tools have zero network/filesystem/OS access by default — all capabilities come through host functions.
- The Gatekeeper is host functions, not a container. For WASM tools, the Gatekeeper logic runs directly inside Electron as host functions gated by
GatekeeperPolicy. No separate proxy or container. - Tools are always low-trust. Even tools written by us get the same restrictions in the Sandbox. Trust is architectural, not reputational.
- Every boundary crossing must be: explicit, Core-mediated, and logged.
- No runtime permission escalation — tools declare permissions in their manifest upfront.
GatekeeperPolicyis the base interface for filtering decisions. WASM host functions call it directly now; a future Docker proxy would call the same policy. Same rules, different plumbing.
Key Architecture Docs
| Doc | When to read |
|---|---|
overview.md |
Before modifying any Core or Sandbox code |
manifest.md |
Before changing tool manifests, permissions, or UI panels |
gatekeeper.md |
Before touching outbound network filtering |
container-communication.md |
Before changing how MosAIc talks to tools |
permissions.md |
Before changing the permission model |
data-model.md |
Before modifying Chronicle, Vault, or data flow |
tool-lifecycle.md |
Before changing tool install, launch, or execution flow |
tool-ui.md |
Before changing how tools render UI blocks |
execution-plan.md |
For the Phase 1 implementation sequence |
implementation-status.md |
To check what's built vs planned |
glossary.md |
For term definitions |
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.
- 3d ago First seen · 212 lines · 2,534 tokens per session scan A 65b7e8c9793d
mosaic-companion AGENTS.md is an instructions file published in the GitHub repository hypercycle-development/mosaic-companion (16 stars, last pushed 5d ago), licensed Apache-2.0. It adds 2,534 tokens to every session, about $0.0127 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 instructions, from other repositories
open-codesign CLAUDE.md
Instructions for OpenCoworkAI/open-codesign, covering claude.md — open codesign, what this project is, hard constraints (do not violate), stack & conventions and frontend stack (locked).
opencove AGENTS.md
AGENTS.md instructions for DeadWaveWave/opencove, covering system prompt: opencove project ai developer agent, 1. core directives & golden rules, 2. decision framework (small vs large), a. small change (fast feedback / 小步快反馈) and b. large change (deep thinking / 慎重对齐).
wesight AGENTS.md
Instructions for freestylefly/wesight, covering claude.md, build and development commands, build production bundle (typescript + vite), lint with eslint and run memory extractor tests (node.js built-in test runner).
limboo CLAUDE.md
Instructions for limboo-ai/limboo, covering claude.md, 1. what is limboo?, 2. tech stack (current), 3. project structure and the three electron contexts (critical mental model).
getvect AGENTS.md
Instructions for craigjmidwinter/getvect, covering agents.md, commands (green on a clean checkout), invariants a pr must not break and process.
open-codesign AGENTS.md
Instructions for OpenCoworkAI/open-codesign, covering agents.md - open codesign, what this project is, hard constraints, ai visibility for web work and current architecture direction.