LeAgent AGENTS.md

A project guide for LeAgent, an open-source desktop system that lets AI agents plan tasks, use tools, build visual workflows, and stream interactive interfaces.

In plain words
What is it for?
Use it when developing or contributing to LeAgent. It helps locate subsystems, understand how the agent runs, and follow the project's development conventions.
Why use it?
It gives coding agents one concise reference for the project's structure, rules, and key components instead of making them search through many files.

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/vixues/leagent/agents-md
Clone the repo
git clone --depth 1 https://github.com/vixues/LeAgent

Made for: Codex, OpenCode.

Per session 6,799 This file is loaded in full into every session.
When invoked 6,799 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 $0.06799 $0.06799
Opus 5 $0.03400 $0.03400
Sonnet 5 $0.01360 $0.01360
Haiku 4.5 $0.00680 $0.00680

Measured 2d ago against content hash 1f5ef53dd74d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

LeAgent 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.

AGENTS.md · 462 lines

How it starts

The opening of the file, as written. The whole thing — 462 lines — stays where its author put it; the contents beside it link to each section on GitHub.

LeAgent Development Guidelines

AI coding assistants: this file is the concise contributor reference. For deep subsystem docs see docs/technical/.

Project Overview

LeAgent is an open-source desktop AI agent stack that doesn't just chat — it gets work done. It fuses three capabilities most agents keep apart: a streaming agent runtime that plans, calls tools, and self-corrects in one think-act loop; agentic visual workflows where the agent designs, runs, and refines ReactFlow DAGs (every tool is automatically a typed node); and a generative UI layer that streams live, interactive interfaces (KPI boards, slide decks, galleries) into the chat. Key pieces: QueryEngine session orchestrator, 100+ domain tools across 13 categories, a versioned Agent SDK, layered prompt management, cognitive three-store agent memory, a relevance-gated context pipeline, a visual workflow builder (ReactFlow) with a first-class game-art asset pipeline, Agent Skills v1.0, MCP, and a declarative YAML rule engine. Runs locally with zero external dependencies by default (SQLite, single process). Providers: DeepSeek, DashScope (Qwen), OpenAI, Anthropic, Azure OpenAI, Ollama, vLLM.

Architecture

LeAgent/
├── backend/
│   ├── leagent/
│   │   ├── agent/        # QueryEngine, controller (compat shim), planner, subagents, recovery
│   │   ├── sdk/          # Versioned public Agent SDK: runtime re-exports, kernel, protocols
│   │   ├── runtime/      # ExecutionRun plane, AgentRuntime, platform/service wiring
│   │   ├── alembic/      # Alembic migration scripts
│   │   ├── api/          # FastAPI routers (v1, incubating v2)
│   │   ├── bootstrap/    # Tool + workflow-node startup
│   │   ├── channels/     # Outbound + IM ingress (DingTalk, Feishu, WeChat Work, Weixin/iLink, console)
│   │   ├── chat_workflow/ # Chat-workflow compilation + embed
│   │   ├── cli/          # Click CLI (23 modules)
│   │   ├── code/         # Code execution layer (sandbox, workspace, runner)
│   │   ├── config/       # pydantic-settings
│   │   ├── context/      # Source-driven, relevance-gated prompt assembly
│   │   ├── cron/         # Scheduled jobs + repository
│   │   ├── db/           # Persistence: async engine, SQLModel models, repositories
│   │   ├── file/         # Unified file layer (primitives, FileService, storage)
│   │   ├── llm/          # LLM service + providers + transport + streaming + media generation
│   │   ├── mcp/          # Model Context Protocol
│   │   ├── memory/       # AgentMemory (episodic/semantic/procedural) + recall
│   │   ├── project/      # Coding project layer (fs, tools, manager, templates)
│   │   ├── prompts/      # PromptBuilder, registry, templates
│   │   ├── rules/        # Rule engine + YAML definitions
│   │   ├── services/     # ServiceManager, auth, chat, session, gen-ui, ...
│   │   ├── skills/       # Agent Skills v1.0
│   │   ├── tasks/        # Background task system + handlers
│   │   ├── telemetry/    # OTel tracing, structured logging, trace propagation
│   │   ├── tools/        # 100+ tools across 13 categories (doc, web, data,
│   │   │                 # db, gen, canvas, chart, image, media, skills,
│   │   │                 # workflow, integration, util)
│   │   └── workflow/     # DAG engine, nodes, art pipeline, templates
│   └── tests/
├── frontend/             # React 19 + TypeScript SPA
├── desktop/              # Electron shell (bundled Python runtime)
├── deploy/               # Dockerfile + SQLite-only Compose
├── config/               # Demo workflows + workflow templates
├── docs/                 # Architecture, guides, deployment docs
├── scripts/              # Install scripts (sh, ps1, bat)
├── website/              # Project website
└── start.sh / start.ps1  # Dev orchestrator (uv + npm)

Read the full file on GitHub · 462 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 · 462 lines · 6,799 tokens per session scan A 1f5ef53dd74d

Subscribe to this mod's changes

LeAgent AGENTS.md is an instructions file published in the GitHub repository vixues/LeAgent (216 stars, last pushed 21d ago), licensed Apache-2.0. It adds 6,799 tokens to every session, about $0.0340 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.