0xeb/ghidrasql-skills

21 stars Quality 57 Trend 24NOASSERTION

Claude Code skills for ghidrasql — live SQL interface for Ghidra analysis and annotation workflows

Overview

Claude Code / Copilot CLI skills for ghidrasql, a live SQL interface for Ghidra analysis and annotation workflows. - A Ghidra project with the LibGhidraHost extension installed and running when you need live queries - ghidrasql available on your PATH or callable from a known location - Verify: ghidrasql --help - For multi-program projects, use --list-project-programs, the project_programs table, or HTTP GET /project/programs, then select one active domain path with --program or HTTP POST /project/open Use the plugin packaging, not a flat copy into ~/.codex/skills. The plugin path preserves the ghidrasql namespace so generic skill names like xrefs, data, and types do not collide with other skills. 2. Copy the plugin directory into your home plugins folder: 3. Create or update ~/.agents/plugins/marketplace.json so it contains: The Codex plugin manifest lives at plugins/ghidrasql/.codex-plugin/plugin.json, and the skills live under plugins/ghidrasql/skills/.

README

ghidrasql-skills

Claude Code / Copilot CLI skills for ghidrasql, a live SQL interface for Ghidra analysis and annotation workflows.

Prerequisites

A Ghidra project with the LibGhidraHost extension installed and running when you need live queries
ghidrasql available on your PATH or callable from a known location
Verify: ghidrasql --help
For multi-program projects, use --list-project-programs, the project_programs table, or HTTP GET /project/programs, then select one active domain path with --program or HTTP POST /project/open

Installation

Claude Code / Copilot CLI

/plugin marketplace add 0xeb/ghidrasql-skills

Codex

Use the plugin packaging, not a flat copy into ~/.codex/skills. The plugin path preserves the ghidrasql namespace so generic skill names like xrefs, data, and types do not collide with other skills.

Home-local install

Clone this repo anywhere temporary:

git clone https://github.com/0xeb/ghidrasql-skills

Copy the plugin directory into your home plugins folder:

mkdir -p ~/plugins
cp -R ghidrasql-skills/plugins/ghidrasql ~/plugins/

Create or update ~/.agents/plugins/marketplace.json so it contains:

{
  "name": "ghidrasql-tools",
  "interface": {
    "displayName": "GhidraSQL Tools"
  },
  "plugins": [
    {
      "name": "ghidrasql",
      "source": {
        "source": "local",
        "path": "./plugins/ghidrasql"
      },
      "policy": {
        "installation": "AVAILABLE",
        "authentication": "ON_INSTALL"
      },
      "category": "Engineering"
    }
  ]
}

Restart Codex.

The Codex plugin manifest lives at plugins/ghidrasql/.codex-plugin/plugin.json, and the skills live under plugins/ghidrasql/skills/.

Repo-local install

If you want to test directly from a checkout, keep the repo where it is and use the bundled marketplace file at .agents/plugins/marketplace.json. Codex should resolve ./plugins/ghidrasql relative to the repo root.

Compatibility

SKILL.md is the canonical contract for each skill.
agents/openai.yaml adds optional Codex app metadata and does not replace SKILL.md.

Skills

Skill	Description	When to Use
`connect`	Bootstrap sessions, REPL, and HTTP usage	Starting work, checking connectivity, routing to the right skill
`analysis`	High-level binary triage and query workflows	Audits, summaries, “what does this binary do?” questions
`annotations`	Persistent renames, comments, signatures, and save workflow	Cleaning up code, adding comments, applying structured annotations
`decompiler`	Pseudocode, locals, parameters, and decompiler-backed workflows	Decompiling functions, local-variable cleanup, read-after-write verification
`disassembly`	Functions, instructions, blocks, CFG, and raw code structure	Instruction-level analysis, block queries, code-shape inspection
`data`	Strings, memory bytes, typed data, and hexdumps	String hunting, data-item recovery, byte inspection
`debugger`	Breakpoints, patch inventory, and debugger-state workflows	Inspecting or changing breakpoints and reviewing patched bytes
`grep`	Name and pattern lookup across symbols and types	Finding functions, names, imports, exports, and type names
`re-source`	Source reconstruction and structure-recovery workflows	Recovering structs, field layouts, and source-level intent
`xrefs`	Callers, callees, call graph, and string-reference tracing	Dependency analysis, import usage, call-chain questions
`types`	Type import, type exploration, and signature application	Structs, enums, typedefs, signatures, and declaration parsing
`functions`	SQL function catalog and function-oriented usage patterns	Looking up `ghidrasql` helper functions and choosing the right one

Support

If ghidrasql and these skills save you time, a donation funds the next release, and a star helps others find them.

Bug reports with a minimal reproduction are just as valuable.

Author

Elias Bachaalany (@0xeb)

License

This project and all its contents — including skill definitions, reference documentation, and configuration files — are licensed under the Human-Origin Source License v1.0.

View the full README on GitHub

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add 0xeb/ghidrasql-skills

New to MCP? Start with the mcpskill.org guide →