ghidra-headless-mcp

Provides a headless Ghidra server that exposes Ghidra reverse-engineering workflows to AI agents via the MCP (Model Context Protocol), with stdio (default) and TCP transports and both real (pyghidra) and fake-backend modes.

Evaluated Mar 30, 2026 (0d ago)
Repo ↗ Ai Ml reverse-engineering ghidra pyghidra mcp headless agent-tools scripting disassembly decompilation patching
⚙ Agent Friendliness
50
/ 100
Can an agent use this?
🔒 Security
31
/ 100
Is it safe for agents?
⚡ Reliability
31
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
78
Documentation
70
Error Messages
0
Auth Simplicity
20
Rate Limits
0

🔒 Security

TLS Enforcement
60
Auth Strength
5
Scope Granularity
0
Dep. Hygiene
50
Secret Handling
50

README explicitly says MCP communication is unauthenticated by default and exposes broad scripting/API access (ghidra.eval / ghidra.call). This is appropriate only for trusted, containerized environments; TCP should not be exposed to untrusted networks. README does not describe TLS, authorization, fine-grained scopes, auditing, or secret-handling specifics.

⚡ Reliability

Uptime/SLA
0
Version Stability
45
Breaking Changes
30
Error Recovery
50
AF Security Reliability

Best When

Run locally or inside a trusted container/VM network with the agent host, and when you can securely manage access to a powerful but unauthenticated scripting-capable MCP interface.

Avoid When

Avoid exposing the TCP transport to untrusted networks/users; avoid using in shared/multi-tenant environments without strong network controls and process isolation.

Use Cases

  • Automated disassembly/decompilation analysis in containerized or sandboxed environments
  • Agent-driven patching and iterative refinement of Ghidra analysis state (types/symbols/xrefs)
  • Running Ghidra-related scripted workflows through MCP tools
  • Integrating reverse-engineering capabilities into an MCP-capable agent host
  • CI/testing of tool catalogs and workflows via fake backend mode

Not For

  • Untrusted network exposure (unauthenticated by default)
  • Use cases requiring production-grade service-level guarantees without additional ops hardening
  • Highly regulated environments without clear compliance/security controls around scripting and remote code execution surfaces

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: none (unauthenticated by default)
OAuth: No Scopes: No

README states MCP communication (stdio/tcp) is unauthenticated by default; no auth mechanism described in documentation.

Pricing

Free tier: No
Requires CC: No

Open-source Python package (license GPL-2.0 per repo metadata). No pricing information provided.

Agent Metadata

Pagination
offset/limit with has_more/next_offset
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • TCP transport is unauthenticated by default; ensure network isolation or use stdio where possible.
  • Tool catalog listing can return full tool set without pagination unless offset/limit are provided (agents may need to cap results).
  • Fake-backend mode is intended for CI/dev without a Ghidra install; capabilities may differ from live mode.
  • Read-only mode for opened programs is default; switching to read-write may be required for mutation workflows.

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for ghidra-headless-mcp.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

6533
Packages Evaluated
19870
Need Evaluation
586
Need Re-evaluation
Community Powered