DEDUPED Hot AI repos. Real use cases. What people build. Get the Monday 5

BEST AI REPOS BY TASK

Best open-source AI repos for securing AI coding agents (2026)

5 repos compared · updated 2026-10-29

What are the best open-source tools to secure AI coding agents and the skills they install?

The best open-source AI repos for securing AI coding agents right now are SkillSpector, SolonGate Agent Security and jev-guard. Scan third-party skills and MCP servers with SkillSpector before you install them, then gate the agent's tool calls: SolonGate Agent Security checks calls against policy and catches secrets, jev-guard returns deny, ask or allow for each call, and Tracekit signs every call and decision into a tamper-evident log you can verify offline.

Comparison

RepoWhat it doesStarsStart here
SkillSpector
NVIDIA/SkillSpector · Python		SkillSpector is NVIDIA's security scanner for AI agent skills used by Claude Code, Codex CLI, Gemini CLI and others.22,109Scan an agent skill for prompt injection and exfiltration before installing Beginner, 10 minutes
SolonGate Agent Security
solongate/agent-security · Go		SolonGate Agent Security is a local, open-source Go gateway that checks AI coding agent tool calls against policy, detects secrets, and writes an on-device audit trail.156Stop an AI coding agent from reading your .env secrets Intermediate, 30 minutes
jev-guard
leepokai/jev-guard · JavaScript		jev-guard is a security hook for coding agents that scores each tool call with session context and returns deny, ask, or allow, flags prompt injection in tool results, and checks skills and plugins, powered by Jev.127Guard Claude Code tool calls with deny and ask decisions Intermediate, 15 minutes
Tracekit
Cygnux-Labs/Tracekit · Python		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.244See a denied command and a tamper check in 5 seconds Beginner, 15 minutes
ForgeGuardian
Mah3Sec/ForgeGuardian · Go		ForgeGuardian is a local-first, Go-based software supply chain security scanner that checks project dependencies across 9 ecosystems with 8 scan engines.n/aScan project dependencies for risks before release Beginner, 15 minutes

When to pick which

SkillSpector NVIDIA/SkillSpector

SkillSpector is NVIDIA's security scanner for AI agent skills used by Claude Code, Codex CLI, Gemini CLI and others. It scans Git repos, URLs, zips, directories or single files for 71 vulnerability patterns in 17 categories, including prompt injection, data exfiltration and MCP tool poisoning, with fast static analysis plus optional LLM evaluation, live OSV.dev CVE lookups, a 0 to 100 risk score and Terminal, JSON, Markdown or SARIF reports.

Best for: Developers and security teams who install third-party agent skills or MCP servers and want to know whether they are safe first.

Needs: Python with uv or pip, or Docker (no Python required); Optional: an LLM API key in .env for semantic analysis (default config targets DeepSeek); Optional MCP extra for skillspector mcp

Start here: Scan an agent skill for prompt injection and exfiltration before installing Beginner, 10 minutes, works with claude-code, codex, gemini-cli

  1. Install with uv tool install git+https://github.com/NVIDIA/skillspector.git
  2. Download or clone the skill you want to check
  3. Scan it without LLM costs using Docker: docker run --rm -v "$PWD:/scan" skillspector scan ./my-skill/ --no-llm
  4. Read the risk score, severity labels and recommendations

All 3 use casesGitHub

SolonGate Agent Security solongate/agent-security

SolonGate Agent Security is a local, open-source Go gateway that checks AI coding agent tool calls against policy, detects secrets, and writes an on-device audit trail.

Best for: Developers and security-minded engineers who run AI coding agents on their own machines and want to control what those agents can read, write, and send.

Needs: A machine where you can run a shell and git (the README installs with git clone and ./scripts/shell/install.sh); Claude Code or Codex installed as the agent you want to guard; A new terminal after install, since hooks load only when a session starts; For the sandbox use case: Linux with bubblewrap (bwrap) installed, since the README says it is required and macOS and Windows refuse

Start here: Stop an AI coding agent from reading your .env secrets Intermediate, 30 minutes, works with claude-code, codex

  1. git clone https://github.com/solongate/agent-security.git
  2. cd agent-security
  3. ./scripts/shell/install.sh
  4. Open a new terminal, then start Claude Code or Codex in your project.

All 4 use casesGitHub

jev-guard leepokai/jev-guard

jev-guard is a security hook for coding agents that scores each tool call with session context and returns deny, ask, or allow, flags prompt injection in tool results, and checks skills and plugins, powered by Jev.

Best for: Developers who run coding agents such as Claude Code, Codex, Cursor or Gemini CLI on real repositories and want an extra approval layer before risky commands run.

Needs: Node.js and npm, to install the jev-guard package or run the install commands; One provider key: a TypeSafe key from console.typesafe.ai, a vck_ Vercel AI Gateway key, or an sk-or- OpenRouter key; The coding agent you want to guard, installed on the same machine

Start here: Guard Claude Code tool calls with deny and ask decisions Intermediate, 15 minutes, works with claude-code

  1. Run npm i -g jev-guard
  2. Run jev-guard key "…" with your TypeSafe, Vercel AI Gateway or OpenRouter key
  3. Run jev-guard install claude
  4. Start Claude Code and run a harmless command, then run jev-guard check Bash '{"command":"rm -rf ~/"}' to confirm the deny output

All 5 use casesGitHub

Tracekit Cygnux-Labs/Tracekit

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.

Best 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.

Needs: Python 3.9 to 3.13 (framework integrations need 3.10 or newer), or uv for the one-line demo; A terminal on your own machine; runs on your own machines per the README; Dev mode runs as your user, so it protects nothing against an agent running as you

Start here: See a denied command and a tamper check in 5 seconds Beginner, 15 minutes

  1. Install uv from https://docs.astral.sh/uv/ if you do not have it.
  2. Run uvx --from tracekit-ai tracekit demo --server.
  3. Read the agent run lines: Bash 'sudo rm -rf /' shows deny, and the first Integrity line says VERIFIED.
  4. Read the tamper test: the rewritten copy shows [FAIL] signatures and Integrity: FAILED at the changed record.

All 4 use casesGitHub

ForgeGuardian Mah3Sec/ForgeGuardian

ForgeGuardian is a local-first, Go-based software supply chain security scanner that checks project dependencies across 9 ecosystems with 8 scan engines.

Best for: Developers and security engineers who need to check a project's dependencies for known CVEs and malicious packages before shipping.

Needs: A terminal (bash on Linux or macOS, or PowerShell on Windows), or Docker; A project directory with dependencies to scan; Git and make only if building from source

Start here: Scan project dependencies for risks before release Beginner, 15 minutes

  1. Install Forge Guardian with curl -sSfL https://raw.githubusercontent.com/Mah3Sec/ForgeGuardian/main/install.sh | bash
  2. Run fgctl update to refresh the community signatures
  3. Change into your project folder and run fgctl scan .
  4. Read the findings, noting severity and the fix version listed for each

All 4 use casesGitHub

Use cases for securing AI coding agents

Questions

What are the best open-source tools to secure AI coding agents and the skills they install?

The best open-source AI repos for securing AI coding agents right now are SkillSpector, SolonGate Agent Security and jev-guard. Scan third-party skills and MCP servers with SkillSpector before you install them, then gate the agent's tool calls: SolonGate Agent Security checks calls against policy and catches secrets, jev-guard returns deny, ask or allow for each call, and Tracekit signs every call and decision into a tamper-evident log you can verify offline.

Which repo for securing AI coding agents is easiest to start with?

SkillSpector: its starter use case "Scan an agent skill for prompt injection and exfiltration before installing" is rated Beginner and takes about 10 minutes.

How do I get started with SkillSpector?

Install with uv tool install git+https://github.com/NVIDIA/skillspector.git; Download or clone the skill you want to check; Scan it without LLM costs using Docker: docker run --rm -v "$PWD:/scan" skillspector scan ./my-skill/ --no-llm. Full steps and a copy-paste agent prompt: https://dedupedai.com/use-cases/nvidia-skillspector/scan-an-agent-skill-before-installing/

Are these securing AI coding agents repos free?

They are open-source projects on GitHub (check each license; some, like fair-code projects, limit commercial use). Some also sell a hosted version or need a paid model API key; the requirements for each repo are listed on this page.

How is this list of securing AI coding agents repos chosen?

Deduped tracks the hottest AI repos on GitHub and writes step-by-step use cases from each README. For this page we picked the repos that do this job, ordered by fit and then GitHub stars. Repos with a risk flag or a retired page are left out. The page is rebuilt with the site; stars are as of the last build in 2026.

More answers

For AI agents: Markdown version · JSON · MCP server · llms.txt

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer