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/lfyxhappy/lfcode/agents-mdgit clone --depth 1 https://github.com/lfyxhappy/lfcodeWhat 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 | $0.03624 | $0.03624 |
| Opus 5 | $0.01812 | $0.01812 |
| Sonnet 5 | $0.00725 | $0.00725 |
| Haiku 4.5 | $0.00362 | $0.00362 |
Grade A, and why
lfcode 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 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.
How it starts
The opening of the file, as written. The whole thing — 210 lines — stays where its author put it; the contents beside it link to each section on GitHub.
- To regenerate the JavaScript SDK, run
bun run --cwd packages/sdk/js build. - The default branch in this repo is
dev. - Local
mainref may not exist; usedevororigin/devfor diffs. - Treat
.codex/*.mdplans as execution-tracking documents, not just notes: before starting plan-scoped work, read the referenced plan and respect its currentStatus:. - If the user asks for a plan first, or the task is being tracked in
.codex/, create or update the relevant plan file before implementation starts and give it an explicitStatus:. - Keep the plan lifecycle synchronized with reality: use
planned/pendingbefore execution, switch toin_progressonce implementation actually starts, and leave it non-completed if work is only partially done or verification is still missing. - When implementing a task from a plan under
.codex/, update that plan in the same change flow: adjust scope/details if they changed, record any remaining gaps, and mark the plan'sStatus:ascompletedonly after the task is fully implemented and verified. - When a
.codex/plan is completed and its changes affect the desktop app, runtime, settings, MCP/tools, or other behavior the user is expected to test in the local use-copy, do not stop at code/test completion: finish the same flow withbun run package:win, sync toC:\算法\小应用\Lfcode, and relaunch the use-copy unless the user explicitly says not to sync. - Keep
.codex/*.mdreserved for plans and execution tracking. Store standalone HTML design mockups, visual replicas, and reviewable static UI drafts under.codex/design-html/instead ofoutput/or scattered temp locations. - When creating a design draft in
.codex/design-html/, prefer a self-contained.htmlfile with inline CSS/assets when practical so it can be opened and reviewed directly without extra build steps.
Project Structure
- This repo is a Bun workspace monorepo rooted at
package.json. - Most product code lives under
packages/. - The core runtime and session engine live in
packages/lfcode. - The web UI lives in
packages/app, and the Electron host lives inpackages/desktop. - Shared packages such as
packages/core,packages/ui,packages/plugin,packages/sdk/js,packages/shared, andpackages/scriptsupport the app/runtime layers. - Cross-repo specs live in
specs/. - User-local project config and content under the repo root belong in
.lfcode/(agent/,command/,skills/,themes/,plugins/,tool/,glossary/, plusopencode.jsoncandtui.json). - App settings surfaces now include
packages/app/src/components/settings-skills.tsxandpackages/app/src/components/settings-archives.tsx; skill-management and archive changes should be validated through those settings views, not only through server routes. - Instance-side skill APIs live under
packages/lfcode/src/server/routes/instance/skills.ts; when changing local skill import, discovery, hiding, or deletion flows, keep the app settings UI and this route in sync.
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.
- yesterday First seen · 210 lines · 3,624 tokens per session scan A 11bedcb03859
lfcode AGENTS.md is an instructions file published in the GitHub repository lfyxhappy/lfcode (2 stars, last pushed 1mo ago), licensed MIT. It adds 3,624 tokens to every session, about $0.0181 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
caelo CLAUDE.md
Instructions for AuraVixStudio/caelo, covering claude.md, what this is, the single most important structural fact, architecture and commands.
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
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.