claude

agentdecompile

PyGhidra-based MCP server for objdiff-verified decompilation recovery

0 on GitHubby AgentDecompilev0.1.0AGPL-3.0-onlyupdated 17 days agodevelopment-tools
/

Plugin agentdecompile from marketplace agentdecompile in bodencrouch/AgentDecompile. Commands read from the repo's manifest — 12 days ago.

What it installs

Skills (17)

  • agentdecompile-rewrite-worker

    Use this skill when a recovery work_dir has pending rewrite-queue entries, or the user wants /loop rewrite fulfillment during an --autonomous campaign. Requires a work_dir.

  • agentdecompile-server-env

    Use this skill when starting the AgentDecompile MCP server, setting GHIDRA_INSTALL_DIR, auto-match, auto-checkin, analysis-tier, or debugging session-id or proxy forwarding, including on Cursor Cloud.

  • bsim-corpus

    Use this skill when ingesting a Ghidra repository into BSim, asking what a bsim_datadir holds, matching names across related builds, or when BSim found nothing might mean createdatabase never ran (issue 169). Apply even if they only say the 24 programs or odyssey repo.

  • lfg

    Use this skill when asked to continue authorized AgentDecompile work from current strategy and evidence.

  • mcp-debugging

    Use this skill when MCP calls time out, tools are missing, schemas mismatch, or the user mentions MCP Inspector, mcptools, mcp-debug, mcp-trace, FastMCP, sandbox, or GUI/coord failures.

  • re-aggregator

    Use this skill when asked to merge several function analyses, resolve disagreements, or report RE consensus and remaining gaps.

  • re-critic

    Use this skill when asked to review, challenge, or verify someone else's function analysis against the binary. Do not invent a new hypothesis.

  • re-planner

    Use this skill when the user wants a plan for a binary, triage of an executable, or the multi-agent RE pipeline. Do not decompile here; dispatch workers.

  • re-worker

    Use this skill when asked to analyze, decompile, or type a single function and return a JSON artifact. Not for coordinating other agents.

  • skill-authoring

    Use this skill when adding, editing, or reviewing AgentDecompile skills or subagents. Apply even if the user only pastes agentskills.io or says skills feel unused. Follow the whole site, not only the specification page.

  • source-command-capabilities

    Use this skill when invoking the migrated AgentDecompile capabilities command.

  • source-command-help

    Use this skill when invoking the migrated AgentDecompile help command.

  • source-command-lfg

    Use this skill when invoking the migrated AgentDecompile lfg command.

  • source-recovery

    Use this skill when the user wants recovered C, objdiff, --dump-source, rematch, or reconstruct. Do not treat leftover target/ receipts as this run's source.

  • tiered-re-analysis

    Use this skill when starting reverse engineering, picking tools for a binary, or before open-project or decompile-function. Apply even if they only say look at this exe or what is in this binary, without naming Ghidra or tiers.

  • web-design-guidelines

    Review UI code for Web Interface Guidelines compliance. Use when asked to "review my UI", "check accessibility", "audit design", "review UX", or "check my site against best practices".

  • workbench-operator

    Use this skill when working in /dashboard, dogfooding the workbench, or making the UI feel like IDA, Ghidra, or x64dbg. Apply even if they only say the dashboard is empty, boilerplate, or unusable.

MCP servers (1)

  • agentdecompile

    connects to: http://127.0.0.1:8080/mcp

View source on GitHub bodencrouch.github.io

Similar plugins

narsil

claude

Code intelligence plugin with 76+ tools for security scanning, call graph analysis, symbol search, and static analysis. Includes slash commands for common workflows and a skill for effective tool usage.

1 skill · 5 commands

postrv/narsil-mcp 187

cc10x

claude

The Loop Engine: a harness for Claude Code. One router sends each request to one of eight workflows (build, debug, review, plan, QA, orient, triage, codebase health) run by specialist subagents, with fail-closed gates, workflow artifacts on disk, hook guardrails, test honesty gates, fresh-context verification, and memory that survives compaction.

22 skills · 14 agents · hooks

romiluz13/cc10x 167

24 audit agents + 6 workflow skills for solo dev startups building with Next.js + Claude Code

6 skills · 24 agents

undeadlist/claude-code-agents 150

universal

claude

Universal AI agents collection: 28+ specialized agents with strict boundary enforcement, project-specific agent creation (/initx), and intelligent task routing (Level 0-5 complexity assessment)

2 commands · 28 agents

Toskysun/sub-agents 111

AI-powered architecture documentation framework with ADRs, reviews, and pragmatic mode

7 skills · 8 agents · 1 MCP server · hooks

codenamev/ai-software-architect 80

Elite Android development skills - ADB mastery, fastboot sorcery, ROM building, and LineageOS development

4 skills · 1 agent

hyperb1iss/hyperdroid-skill 55

Maintain agentdecompile?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted