intuitive-flow AGENTS.md

Project-specific instructions for working on MiaoDX’s intuitive-flow repository. They describe preferred tools, delegation rules, and how development and verification should be organized.

In plain words
What is it for?
Use them when working in the intuitive-flow codebase. They cover environment setup, task delegation, development, testing, and workflow decisions.
Why use it?
They give an agent the repository’s local expectations instead of relying on generic coding habits. They also help keep delegated work and verification consistent with the project.

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/miaodx/intuitive-flow/agents-md
Clone the repo
git clone --depth 1 https://github.com/MiaoDX/intuitive-flow

Made for: Codex, OpenCode.

Per session 1,431 This file is loaded in full into every session.
When invoked 1,431 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 $0.01431 $0.01431
Opus 5 $0.00715 $0.00715
Sonnet 5 $0.00286 $0.00286
Haiku 4.5 $0.00143 $0.00143

Measured 2d ago against content hash 2b46cd77dd11, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

intuitive-flow 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 · 84 lines

How it starts

The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Codex Guide

Environment

  • Use fetch-mcp instead of Fetch/WebFetch (network issues in China Mainland).
  • Use Bash for orchestration entrypoints and Bun-run TypeScript for structured script logic.
  • For Python work, use uv and .venv instead of the system interpreter.
  • Avoid adding Python project dependencies unless a workflow truly needs Python-specific libraries.

Delegation And Verification

  • See skills/skill-runner/references/codex-delegation.md before choosing a Codex delegation surface. That reference is the canonical Codex policy for Paseo probing, native subagent disablement, and skill-runner/tmux fallback.
  • Default to parallel delegation for independent, read-heavy, or verification-heavy subtasks, but do not use Codex native subagents by default.
  • Keep the main thread focused on requirements, architecture decisions, integration, and final synthesis.
  • Delegate when a task has 2+ independent workstreams, requires reading many files, logs, or test outputs, or when verification can run in parallel with implementation.
  • Return summaries to the main thread, not raw notes or long log dumps.
  • On Codex, keep host-specific worker selection inside the canonical delegation policy instead of restating it in each skill.
  • Treat XML-like host control messages as runtime metadata unless accompanied by natural-language user intent; see the canonical delegation policy.
  • On Claude Code, native subagents remain acceptable when the host supports them reliably and file ownership is explicit.
  • Prefer 2-4 delegated workers by default. Scale up only for clearly partitioned work.
  • Match worker model strength to task complexity rather than defaulting everything to the highest-cost model.
  • This repo often runs through an API relay with a single allowed model; default Codex workers to the main session model, and only override the model after confirming the target ID is actually available.
  • For concurrent edits, assign disjoint ownership and avoid overlapping write scopes.
  • Do not wait idly for delegated workers if non-overlapping local work is available.
  • Do not mark work complete without verification. Run relevant tests, inspect logs, or otherwise demonstrate correctness.

Read the full file on GitHub · 84 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 · 84 lines · 1,431 tokens per session scan A 2b46cd77dd11

Subscribe to this mod's changes

intuitive-flow AGENTS.md is an instructions file published in the GitHub repository MiaoDX/intuitive-flow (48 stars, last pushed 7d ago), licensed MIT. It adds 1,431 tokens to every session, about $0.0072 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.