Ghidra MCP’s star growth jumps as version 7 enters release-candidate testing
The Apache-2.0 reverse-engineering bridge accelerated from roughly 46 to 209 GitHub stars per day across comparable five-day windows, while version 7 remains a testing-only pre-release.
2 min read
Ghidra MCP, a community-maintained bridge between AI agents and the Ghidra reverse-engineering platform, is drawing sharply more GitHub attention as its next major version moves through release-candidate testing.
The canonical bethington/ghidra-mcp repository showed about 5.4K stars and 283 forks when checked on October 11. A tracker with 64 dated snapshots recorded 5,339 stars the same day. In its directly comparable five-day windows, the project gained 1,043 stars from October 6–11—about 209 per day—versus 228 stars, or roughly 46 per day, from October 1–6. That is a measured acceleration of about 4.5×, not an estimate from a trending-list position. Stars measure attention, however, not production adoption or software quality.
Maintained under the bethington account and licensed under Apache 2.0, Ghidra MCP lets clients such as coding agents call Ghidra through Model Context Protocol. Its current documentation covers GUI and headless operation, Docker deployment, batch analysis, decompilation, call graphs, P-code emulation, live debugging, shared Ghidra Server workflows and write operations such as renaming or adding types and comments.
GitHub still marks version 6.0.0 as the latest stable release. Version 7.0.0-rc.2 was published October 9 as a testing-only pre-release for Ghidra 12.1.4, listing 209 MCP tools and fixes to release checks, debugger proxy registration, scripts and documentation. The preceding rc.1 listed 253 tools, so users should not assume tool counts are directly comparable without checking the shipped catalog.
The breadth also raises operational risk. Inline Ghidra scripts can execute arbitrary Java inside the process and are disabled by default; enabling them requires GHIDRA_MCP_ALLOW_SCRIPTS=1. Users should keep the bridge local, restrict it to authorized binaries and test the release candidate separately from stable workflows.
Sources: Canonical repository · Release history · Dated star history · Weekly growth methodology · Unsplash license
Conversation
0 approved commentsNo comments yet. Start a thoughtful conversation.
Join the conversation.
Sign in with ChatGPT to leave a comment or rate this article.