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/osama-raddad/firecrasher/agents-mdgit clone --depth 1 https://github.com/osama-raddad/FireCrasherWrote 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/osama-raddad/firecrasher/agents-md)<a href="https://agentmods.dev/instructions/osama-raddad/firecrasher/agents-md"><img src="https://agentmods.dev/badge/instructions/osama-raddad/firecrasher/agents-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 5.1 | $0.01780 | $0.01780 |
| Opus 5 | $0.00890 | $0.00890 |
| Sonnet 5 | $0.00356 | $0.00356 |
| Haiku 4.5 | $0.00178 | $0.00178 |
Grade A, and why
FireCrasher 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 6d 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 — 135 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents working in this repository. This is the
canonical instructions file; CLAUDE.md points here. Module-specific notes
live in firecrasher/AGENTS.md and app/AGENTS.md.
What this is
FireCrasher is an Android library that catches uncaught exceptions and
runs a staged recovery sequence to keep the app alive instead of letting
it die. It is published to JitPack as com.github.osama-raddad:FireCrasher.
Current version is 3.0.0.
The recovery escalation ladder has three levels (RecoveryLevel):
- RESTART_ACTIVITY — restart the crashed activity (
recreate()on the first attempt, then relaunch-and-finish). Used whileretryCount <= 1. - GO_BACK — the activity keeps crashing, so pop the back stack (go back) if there is another activity behind it.
- RELAUNCH_APP — nothing to go back to, so relaunch the whole app from its launcher activity.
The core trick (FireLooper) is to replace the main-thread message loop with a
reflection-driven loop wrapped in a try/catch, so an uncaught exception on the
main thread does not tear down the process — the loop reposts itself and
recovery runs.
Module layout
This is a two-module Gradle project (settings.gradle):
firecrasher/— the published library. This is where real work happens.app/— a demo application (com.osama.firecrasherdemo) that installs FireCrasher and deliberately crashes to exercise recovery. Not published.
Library source (firecrasher/src/main/java/com/osama/firecrasher/)
| File | Role |
|---|---|
FireCrasher.kt |
Public API entry point: the FireCrasher object (install), the Application.installFireCrasher extension, and the internal recovery policy (evaluate, recover, dispatchCrash, retryCount, recovery-state persistence). |
FireCrasherConfig.kt |
The @FireCrasherDsl marker and the install DSL builder: onCrash { } (CrashScope receiver, default { recover() }) and onPreviousProcessExit { }. |
CrashScope.kt |
Receiver of the onCrash handler: throwable, activity, level, retryCount, and recover(level = evaluated) { }. |
RecoveryLevel.kt |
The three-level recovery enum (RESTART_ACTIVITY, GO_BACK, RELAUNCH_APP). Declaration order is persisted by the codec — append only. |
ExitInfo.kt |
Public Context extensions historicalExitReasons(maxCount) / lastAbnormalExit() (API 30+, degrade to empty/null below). |
CrashHandler.kt |
Internal Thread.UncaughtExceptionHandler + ActivityLifecycleCallbacks. Tracks the current activity and the live activity count (backStackCount). Delivers crashes to dispatchCrash on the main thread. |
FireLooper.kt |
Reflection-based replacement main loop that survives exceptions. isSafe guards against double-install. |
RecoveryState.kt |
RecoveryState data class + RecoveryStateCodec — encodes recovery progress into a ≤128-byte blob stored via ActivityManager.setProcessStateSummary so recovery escalation survives process death (API 30+). |
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.
- 6d ago First seen · 135 lines · 1,780 tokens per session scan A 0a77f775ebab
FireCrasher AGENTS.md is an instructions file published in the GitHub repository osama-raddad/FireCrasher (148 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 1,780 tokens to every session, about $0.0089 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
sceneview copilot-instructions.md
Copilot instructions for sceneview/sceneview, covering sceneview — github copilot instructions, when to recommend sceneview, android code generation rules, minimal 3d template and minimal ar template.
androidbible CLAUDE.md
Claude Code instructions for yukuku/androidbible, covering claude.md, project overview, branches, build commands and debug apk (open-source build, works out of the box).
SeekerClaw CLAUDE.md
Claude Code instructions for sepivip/SeekerClaw, covering claude.md — seekerclaw project guide, project.md — source of truth, design principle: ux first, what is this project and supported devices.
sceneview CLAUDE.md
Claude Code instructions for sceneview/sceneview, covering sceneview, modules, things that break if you don't know them and conventions.
debroid AGENTS.md
Instructions for PatilShreyas/debroid, covering ai agent contributor guide 🤖, 🎯 intention & motivation, 🏗️ architecture & layers, 1. the core layer (jdi & adb orchestration) and 2. the cli layer (daemon & client).
quark-engine CLAUDE.md
Claude Code instructions for ev-flow/quark-engine, covering claude.md, what is quark engine, commands, install for development and or with pipenv.