Phantom KV alternatives

Phantom KV (lordx64/phantom-kv): phantom-kv is a Python project that tests a refusal-removal method, which loads a small trained bank of key and value tensors into a language model's KV cache without changing the model weights. See its 3 use cases.

Similar repos

1. Rea morluto/rea · TypeScript
   A plug-in that lets your AI agent take apart an app and explain how a feature works. For developers and security-minded builders who want an AI agent to take apart software they can run locally and explain how a feature works, with evidence.
   5 use cases, for example Rebuild a feature you saw, in your own project
2. Reverify 2akouwu/reverify · Python · 1,308 stars
   Reverify is a Python toolkit and MCP server where deterministic tools verify every AI claim about a binary or code before it counts as fact. Reverse engineers, CTF and malware analysts, and developers who use AI coding agents and need their claims checked.
   4 use cases, for example Check an AI's claims about a binary before trusting them
3. Ablation Ablation-Tool/ablation · Python · 222 stars
   Ablation is a self-contained reverse engineering framework that, when used with Codex or Claude Code, works as an autonomous reverse engineering system. Security researchers and reverse engineers who want an AI coding agent to run binary and source analysis modules on firmware, native binaries, or source repositories.
   4 use cases, for example Trace network input to dangerous sinks in a binary
4. Claude AD ADScanPro/Claude-AD · 213 stars
   Claude-AD is a Claude Code plugin that gives Claude an Active Directory pentest method: skills, agents and slash commands for authorized internal assessments that drive standard tools such as netexec, impacket, certipy, bloodyAD and BloodHound CE. Written for penetration testers and red-team operators who hold written authorization for an internal Active Directory assessment and want Claude to follow a structured method.
   4 use cases, for example Scope an AD assessment in the right phase order
5. Awesome Reverse Engineered alexgaoth/awesome-reverse-engineered · Python · 118 stars
   A searchable index of 12,869 open-source tools that were built by reverse engineering closed apps, APIs, devices, protocols, file formats, games and AI products. Developers, security researchers and AI agents that need to check whether a reverse engineered tool already exists before they start new work.
   4 use cases, for example Check for an existing tool before reverse engineering a device
6. Seep Reverse Lab angusdevgo/Seep-Reverse-Lab · Lua · 1,731 stars
   Seep Reverse Lab is a Lua-based, agent-driven workbench that wraps reverse engineering tools (Radare2, JADX, Apktool, Frida) and runs CWE-602 client-side authorization audits. Written for security auditors and reverse engineers who work with an AI agent and need authorized, evidence-backed checks of client-side app behavior.
   4 use cases, for example Audit an app feature gate for CWE-602 authorization flaws
7. Geiger Atomburstofficial/geiger · JavaScript · 167 stars
   Geiger is a read-only, dependency-free command that inventories the AI agents, MCP servers, plugins, and AI extensions on a machine and labels what each one can execute, read, or reach. Developers, IT staff, and security-minded individuals who want to know what AI tools are installed on a machine and what they can touch before trusting them.
   5 use cases, for example See every AI agent and MCP server on a machine
8. Tracekit Cygnux-Labs/Tracekit · Python · 138 stars
   Tracekit is a Python tool that gates each AI agent tool call before it runs, signs every call and decision into a tamper-evident hash chain, and lets anyone verify each run offline. Built for developers and security or audit reviewers who need to see what an AI agent did, whether it was allowed, and whether the record can be trusted.
   4 use cases, for example See a denied command and a tamper check in 5 seconds

Best AI repos for Security and science · All use cases

Questions

Third-party repos we don't own or vouch for. Check each license and README before you run anything. Disclaimer