Android Reverse Engineering Mcp

Safe Android Reverser MCP

**MCP-first, sandboxed Android reverse engineering with semantic program understanding for AI agents.**

Safe Android Reverser is an AI-native reverse code-intelligence platform. The agent asks semantic questions; one host control plane selects and verifies isolated capability workers; analyzers return bounded evidence with provenance.

**The agent reasons. The MCP control plane controls. Capability workers execute.**

**Detect the framework first, then analyze the representation that actually contains the business logic.**

Documentation

• Install / update / troubleshoot
• Project direction
• Roadmap
• Capability SPI v1
• Development and review rules
• Release procedure
• Program-understanding phase 1
• Research: AI-agent program understanding

Coding agents must also follow [`AGENTS.md`](AGENTS.md).

Release status

Release candidate: **0.3.0 — Platform Foundation + Flutter AOT**.

The 0.3 platform architecture has passed exact-head static-core, framework-flutter, and one-public-MCP integration CI and has been merged into `master`. Publication remains gated on the exact 0.3.0 release commit, controlled Flutter runtime-cache verification, immutable semver capability-image publication, and `safe-v0.3.0` release checks.

Quick start

1. Install rootless Podman or Docker
2. /plugin marketplace add salingnh/android-reverse-engineering-mcp
3. /plugin install safe-android-reverser@salingnh-reverse-tools
4. /reload-plugins
5. /mcp
6. Call health
7. Call fingerprint on APK/XAPK/APKS/APKM
8. Follow analysis_route.primary_capability_id

Normal users do not manually clone the repository, register multiple MCP servers, or run analyzer containers themselves.

0.