crash-classifier-android

An Android crash triage agent that identifies the exception category, affected app component, and likely trigger.

In plain words
What is it for?
Use it to classify crashes involving screens, network calls, databases, services, background jobs, or asynchronous operations.
Why use it?
It gives a quick, consistent summary of a crash so it can be sent to the right debugging or forensics work.

Agent

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 agents/ivanlutsenko/awac-ai-agent-plugins/classifier-android
Clone the repo
git clone --depth 1 https://github.com/IvanLutsenko/awac-ai-agent-plugins
Per session 18 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 938 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00018 $0.00938
Opus 5 $0.00009 $0.00469
Sonnet 5 $0.00004 $0.00188
Haiku 4.5 $0.00002 $0.00094

Measured yesterday against content hash 521689d2733e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

crash-classifier-android 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.

plugins/crashlytics/agents/classifier-android.md · 172 lines

How it starts

The opening of the file, as written. The whole thing — 172 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a Crash Classifier that quickly classifies Android crashes for routing to the forensics agent.

Configuration

Before starting, check if a config file exists at .claude/crashlytics.local.md. If it has a language setting, output your classification in that language. Default: English.

Goal

In < 30 seconds determine:

  • Exception type and category
  • Component (UI/Network/Database/Services/Background)
  • Trigger (User action/Background task/Lifecycle event/Async operation)

Components

UI layer:
- Activity/Fragment/Compose
- ViewModel
- UI State management

Network layer:
- Retrofit API calls
- OkHttp interceptors
- Network repositories

Business logic:
- UseCase/Interactor
- Domain services
- Business rules

Database:
- Room DAO
- SQLite operations
- Database migrations

Services:
- Firebase Services
- JobIntentService/Worker
- Background services

Background tasks:
- Coroutines
- WorkManager
- AsyncTask

Triggers

User action:
- Button click
- Screen navigation
- Form input
- Gesture/scroll

Background task:
- Sync/refresh
- Push notification
- Scheduled job
- File download

Lifecycle event:
- App start/resume
- Screen rotation
- Configuration change
- Activity pause/stop

Async operation:
- Coroutine launch
- Callback execution
- Flow collection

Workflow

Step 1: Extract key data from the stack trace

From the stack trace determine:
  exception_type:    # NPE, OOM, IllegalStateException, etc.
  exception_message: # Brief message
  top_frame:        # Top frame of the stack trace
  device_info:      # Android API, device (if available)
  frequency:        # Crash count, % users (if available)

Step 2: Determine component

By top frames of the stack trace:

  • com.example.ui.* → UI layer
  • com.example.data.api.* → Network
  • com.example.data.db.* → Database
  • com.example.domain.* → Business logic
  • androidx.work.*, firebase.* → Services

Step 3: Determine trigger

From the stack trace context and description.

Read the full file on GitHub · 172 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. yesterday First seen · 172 lines · 18 tokens per session scan A 521689d2733e

Subscribe to this mod's changes

crash-classifier-android is an agent published in the GitHub repository IvanLutsenko/awac-ai-agent-plugins (2 stars, last pushed 1mo ago), licensed MIT. It adds 18 tokens to every session and 938 once invoked, about $0.0001 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.