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/jpantsjoha/ai-native-developer-experience/gemini-mdgit clone --depth 1 https://github.com/jpantsjoha/ai-native-developer-experienceWrote 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/instructions/jpantsjoha/ai-native-developer-experience/gemini-md)<a href="https://agentmods.dev/instructions/jpantsjoha/ai-native-developer-experience/gemini-md"><img src="https://agentmods.dev/badge/instructions/jpantsjoha/ai-native-developer-experience/gemini-md.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 | $0.00721 | $0.00721 |
| Opus 5 | $0.00360 | $0.00360 |
| Sonnet 5 | $0.00144 | $0.00144 |
| Haiku 4.5 | $0.00072 | $0.00072 |
Grade A, and why
ai-native-developer-experience GEMINI.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 4d 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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
GEMINI.md — Always-On Harness Rules
The model is the easy part. These rules are the harness. Read them before you generate anything. Speed is easy. Safe speed is engineered.
This file is the always-on "Rules" primitive for this repo. It is portable and opinionated by design — borrow it, adjust it to your team, your mileage may vary.
Design before generate
- Restate the intent before writing code. If you cannot state the goal in one sentence, you are not ready to generate.
- Draft the spec and a first cut together, then review the spec harder than the code. A loose spec is where confident-but-wrong output comes from.
- No speculative scaffolding. Build what the task needs, not what it might need later.
The Adversarial Gate (required reasoning step)
Before you call any change done, answer in writing: "How would I break this?"
- Name the failure modes, the missing guardrail, the untested edge, the input you did not validate.
- Argue against your own approach before you proceed. If you cannot find a way to break it, you have not looked hard enough.
- This gate runs on every non-trivial change. It is not optional.
Receipts, not polish
- Done means checks pass, not that the prose reads well. Evidence over aspiration.
- Every non-trivial change leaves one runnable check behind — a test, an assertion, a command whose output you can paste. If it is not verified, it is not done.
- "Ensure quality" is decoration. "These three checks pass" is a harness.
File standards
- Every source file opens with a short header or module docstring: what it does, and why it exists.
- Public functions get a docstring — inputs, outputs, and the one edge case that bites.
- Match the surrounding file. Reuse the helpers and patterns already in this repo before writing new ones.
- British spelling in prose; follow the language's own convention in code.
Where the harness lives
- Skills / workflows →
skills/(reusable SKILL.md workflows). Look here before inventing a new approach. This is the canonical real directory and the Agent Plugins fixed discovery location;.agents/skills/is a symlinked alias kept for runners that discover there natively. - MCP servers / data seams →
.agents/mcp_config.json(governed access to data sources — a template, not live config; the pattern is documented inDEVELOPER_EXPERIENCE.md). - Portable packaging →
plugin.jsonat the root is the entry point for any Agent Plugins 1.0.0 client. The vendor manifests are projections of it, andmake spec-conformanceproves they have not drifted. - The full guide →
DEVELOPER_EXPERIENCE.md(DX-001): guardrails, gates, spec-driven delivery, CI.
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.
- 4d ago First seen · 49 lines · 721 tokens per session scan A 24441aa7c4de
ai-native-developer-experience GEMINI.md is an instructions file published in the GitHub repository jpantsjoha/ai-native-developer-experience (11 stars, last pushed 27d ago), licensed Apache-2.0. It adds 721 tokens to every session, about $0.0036 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
sprawling CLAUDE.md
Instructions for 2youg1/sprawling: See AGENTS.md. It is the whole instruction set for this repository — the loop, what to read first, the five steps of one change, the rules a machine holds, and the language and commit conventions.
hyper-waterfall CLAUDE.md
Instructions for postmelee/hyper-waterfall: 본 저장소의 에이전트 운영 규칙은 AGENTS.md를 단일 진실 원천으로 한다. Claude Code도 이 규칙을 그대로 적용한다.
mumei CLAUDE.md
Instructions for iroha924/mumei, covering mumei development guide, what mumei is, language policy, bash + jq conventions (essentials) and schemas/ conventions (essentials).
gru-ai CLAUDE.md
Instructions for andrew-yangy/gru-ai, covering gruai — claude code rules, what this is, context tree structure, how to read the context tree and key conventions.
harness-evolver CLAUDE.md
Instructions for raphaelchristi/harness-evolver, covering claude.md, what this is, dependencies, running tools locally and use --no-canary to skip preflight check.
wtcraft CLAUDE.md
Instructions for zywkloo/wtcraft, covering wtcraft routing, repo structure (wtcraft dogfooding), model knowledge policy and release guardrails.