GitHub · by dagoSte

REA Safe Agent: contained REA analysis with Claude Code and Codex

A skill that runs REA, Ghidra and JADX with Claude Code or Codex inside restricted Docker containers. It has no network, shares only hash-approved copies of selected files, and includes repeatable boundary checks.

● Integrations & MCPX posts
REA added as a /reverse-engineer skill in Computer
Announcement that REA (Reverse Engineer Anything) is now a skill in Computer, which inspects how an app works without its source code and then builds a working version.
Host
Computer
Skill
/reverse-engineer
Free through
October 31
Computer
@AskPerplexity
● Integrations & MCPGitHub
ReaFuse
Analyzes two projects with REA, proposes fused ideas, then builds and tests a new project with a copy check.
sickpyramid
@sickpyramid
● Integrations & MCPGitHub
rea-mcp (LM Studio plugin)
Thin wrapper that exposes REA's MCP tools inside LM Studio.
Nowng
@Nowng
● Integrations & MCPGitHub
rea-agent-docker
Docker images for REA, one per disassembler (Ghidra, Hopper, IDA), with a verification harness.
alexmi256
@alexmi256
● Integrations & MCPGitHub
ERA: an analysis workbench and agent workflow layer built on REA
Open-source agent and desktop/browser Studio built on REA. It routes targets to REA workflows, records executions, normalizes evidence into reports, and supports multi-agent review and rebuild in isolated copies.
Version
v0.2.0
REA version
6.4.0
Validated platform
macOS arm64
dav1dngy
github.com

REA Safe Agent: contained REA analysis with Claude Code and Codex

More like this

REA added as a /reverse-engineer skill in Computer

ReaFuse

rea-mcp (LM Studio plugin)

rea-agent-docker

ERA: an analysis workbench and agent workflow layer built on REA