android-native-init-lab AGENTS.md

Repository instructions for an Android project with strict rules for selecting targets, using devices, preserving evidence, and validating changes. They define the required work cycle from checking the current state through reporting and committing.

In plain words
What is it for?
Use them when modifying, testing, or documenting the Android project, especially when deciding whether device testing is required.
Why use it?
They prevent changes from being tested on the wrong target and limit unnecessary or unsafe device actions. They also establish which policies take precedence.

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/temmie-tem/android-native-init-lab/agents-md
Clone the repo
git clone --depth 1 https://github.com/Temmie-Tem/android-native-init-lab

Made for: Codex, OpenCode.

Per session 4,186 This file is loaded in full into every session.
When invoked 4,186 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 $0.04186 $0.04186
Opus 5 $0.02093 $0.02093
Sonnet 5 $0.00837 $0.00837
Haiku 4.5 $0.00419 $0.00419

Measured 2d ago against content hash 694d3bc7a353, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

android-native-init-lab 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 · 261 lines

How it starts

The opening of the file, as written. The whole thing — 261 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md - repository operating contract

Contract-Revision: 2 (supersedes revision 1; 2026-08-03)

The retired Interim Fast-Loop trial contract is preserved byte-for-byte at docs/archive/policy/AGENTS_INTERIM_FAST_LOOP_RETIRED_2026-08-03.md; it is historical evidence only and grants no current authority.


This file contains the repository-wide invariants and the binding target registry. Select exactly one target contract before target-specific work. A registry goal describes current state and objectives but never grants device authority. Historical or draft policies under docs/archive/ or elsewhere are evidence only, even if their text says ACTIVE.

The default work cycle is:

STATE -> SELECT -> DESIGN -> IMPLEMENT -> STATIC VALIDATE -> DEVICE -> REPORT -> COMMIT

Do not add a device step when host-only work can answer the question.

Authority and Precedence

The effective contract is, in descending order:

  1. the common invariants in this file;
  2. the selected binding target contract in the registry below;
  3. the shared risk-tier and execution-process documents named by that target;
  4. the immutable live binding required by the active policy or current runner.

The more restrictive applicable rule wins. A target contract may specialize only behavior explicitly delegated by this file and may never relax the permanent safety boundaries. A manifest, approval string, goal, report, archived clause, helper string, or sub-goal cannot override a higher layer.

No document grants standing device authority unless this common contract or the selected target contract expressly activates it and all required live inputs are current. An unactivated policy edit remains H0 only.

Binding Target Registry

Target Current state Binding target contract Binding live process
Samsung Galaxy S22+ FYG8 (SM-S906N / g0q / S906NKSS7FYG8) GOAL.md docs/operations/targets/S22PLUS_FYG8_TARGET_CONTRACT.md docs/operations/DEVICE_ACTION_PROCESS_V2.md
Samsung Galaxy A90 5G GOAL_A90.md docs/operations/targets/A90_TARGET_CONTRACT.md docs/operations/targets/A90_TARGET_CONTRACT.md sections A90 D1 Resident Session, A90 F1 Resident Install, and Attended F1 Pre-Handoff
Samsung Galaxy S20+ 5G (SM-G986N / y2q / G986NKSS8IYC2) GOAL_S20PLUS.md docs/operations/targets/S20PLUS_G986N_TARGET_CONTRACT.md Active exact-target routine D0/D1 including payload-free Download return; attended boot-only bootstrap and resident Magisk F1 active; reviewed attended native-canary R1 active

Read the full file on GitHub · 261 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 · 261 lines · 4,186 tokens per session scan A 694d3bc7a353

Subscribe to this mod's changes

android-native-init-lab AGENTS.md is an instructions file published in the GitHub repository Temmie-Tem/android-native-init-lab (10 stars, last pushed 3d ago), licensed MIT. It adds 4,186 tokens to every session, about $0.0209 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.