security MCP Server
Agentic bug bounty and pentest orchestrator. Claude Code plugin. Phase-aligned playbooks (recon -> scan -> manual -> chain -> report), parallel sub-agent dispatch, Kali via kali-ssh MCP.
Discovered via github-topic:mcp and last synced 2mo ago.
Install instructions not detected yet
Check the source repository for the latest setup steps.
Trigger phrases
DNS enum + subdomain takeover + cloud asset discovery
"learn mode", "update playbooks", "what's new", "scrape sources"
>---[ 6C 65 74 68 61 6E 69 ]-----------> every move, deliberate ``` [](LICENSE) [](https://claude.com/claude-code) [](#) [](https://www.kali.org) [](examples/demo.gif) <br> <img src="examples/demo.gif" alt="lethani demo" width="820"> <!-- Optionally also upload to asciinema.org and add: [](https://asciinema.org/a/REC_ID) --> <!-- After you push to GitHub, replace USER below with your handle and uncomment: [](https://github.com/onurogut/lethani/commits/main) --> </div> > *"The Lethani is a path one walks. It is the right action moving from the > right heart, the right mind."* — Vashet, *The Wise Man's Fear* ### Why the name? In Patrick Rothfuss's *Kingkiller Chronicle*, the Adem mercenaries follow the **Lethani** — not a moral code, exactly, but a felt sense of *the right action* in any given moment. They are paid killers, yet every motion is deliberate; nothing is wasted; nothing is reckless. That is the discipline this tool tries to inherit. Offensive security work, done well, is the same: authorized, deliberate, evidenced, and never gratuitous. lethani is named for that posture, and tries to enforce it in its defaults — confirm scope by writing it down, fan work out only where it helps, surface critical findings the moment they exist, and stop at the first irreversible operator-side action. --- An **agentic bug bounty and penetration testing orchestrator** packaged as a Claude Code plugin. lethani turns ad-hoc engagement work into a phase-aligned, playbook-driven workflow that fans parallelizable steps out to sub-agents, keeps the context window under 40K tokens, and treats authorization as implicit — it is a tool, not a junior consultant. --- ## What lethani does - **Phase-aligned playbooks** — 51 markdown playbooks covering recon, automated scanning, manual vuln testing, vulnerability chaining, and reporting. Each is a deterministic procedure with steps, output format, and pass/fail criteria. - **Agentic by default** — Phase 1 (14 independent sources) fans out to 6 parallel sub-agents; Phase 2 to 3; Phase 3 buckets endpoints across N. Phases 4 and 5 are single-threaded synthesis. - **Kali-as-execution-plane** — every active command runs on Kali Linux via the `kali-ssh` MCP server. The host edits files and brokers decisions; it does not run scans directly. - **Authorization is implicit** — when you invoke lethani, you have already decided the target is authorized. lethani does not re-ask before each kali-ssh call or before Phase 1. Confirmation is reserved for irreversible operations on your own machine (`rm -rf`, force-push, deleting engagement dirs). - **Context budget enforced** — < 40K tokens per session. Detailed reference lives in `00_infra/` and loads only on demand. - **Learning Mode** — pulls fresh techniques from curated public sources (PortSwigger, watchTowr, Assetnote, HackTricks, PayloadsAllTheThings, HackerOne hacktivity, nuclei-templates, CISA KEV, …), filters by a hard quality bar, and proposes playbook patches that require your approval before any file is written. - **Anonymous reporting** — reports use `Tester : [redacted]` by default. --- ## Install ### Option A — as a Claude Code plugin (recommended) Inside Claude Code, type these two commands one at a time. **The line must start with `/` (no leading space)** — Claude Code only treats the message as a slash command when `/` is the first character. If you paste with a leading space the prompt becomes a regular chat message and the agent replies "I can't run built-in commands." ``` /plugin marketplace add onurogut/lethani ``` ``` /plugin install lethani@lethani ``` The first command registers this repo as a Claude Code plugin marketplace (`.claude-plugin/marketplace.json` is the catalog). The second installs the `lethani` plugin from that marketplace, exposing the slash commands (`/new-target`, `/recon`, `/scan`, …), the 6 recon/OSINT sub-agents, and the 3 skills in any Claude Code session. > If `/plugin install` fails with `Permission denied (publickey)`, Claude > Code is trying to clone via SSH. The repo is public, so HTTPS is fine: > > ```bash > git config --global url."https://github.com/".insteadOf "[email protected]:" > ``` > > Then retry `/plugin install lethani@lethani`. ### Option B — as a workspace ```bash git clone https://github.com/onurogut/lethani.git ~/lethani cd ~/lethani/plugin claude ``` Use this when you want the engagement directory, behavior rules (`plugin/00_infra/behavior_rules.md`), and the CLAUDE.md router all loaded in the session context, not just the plugin commands. Engagements live one level up at `~/lethani/engagements/`. Both modes coexist — installing the plugin and `cd ~/lethani/plugin` in the same session gives you the slash commands plus the workspace-level routing. --- ## First-run setup (Kali + MCP) ```bash # host-side: SSH key, ~/.ssh/config block, MCP sample config ./plugin/00_infra/scripts/setup-host.sh # kali-side: install the tool inventory the playbooks expect scp plugin/00_infra/scripts/setup-kali.sh talon-kali:/tmp/ ssh talon-kali "bash /tmp/setup-kali.sh" ``` Full details and troubleshooting: [SETUP.md](SETUP.md). --- ## First engagement In Claude Code: ``` /new-target acme.com /recon acme ``` lethani will: 1. Scaffold `engagements/acme/` with a `scope.md` template (Phase 0). 2. Fan out Phase 1 to 6 parallel sub-agents (DNS, HTTP, params/CICD, network OSINT, people OSINT, source OSINT). 3. Drop findings into `engagements/acme/recon/` and a summary in `recon/_summary.md`. It will **not** ask "is this in scope?" before running. You created the engagement; that is the authorization record (`behavior_rules.md` §1). Continue through phases: ``` /scan acme # Phase 2 — cve-match + nuclei + ffuf /manual-test acme # Phase 3 — bucketed manual testing /chain acme # Phase 4 — vulnerability chaining (single-threaded) /report acme # Phase 5 — duplicate check + severity + report ``` --- ## Plugin commands
GitHub dorking for secrets/configs
Job
"parallel", "fan out", "sub-agents", "agentic", "run phase N"
tech fingerprint + vhost + JS extraction + wayback
"new target", "test <domain>", "start engagement"
parameter discovery + CI/CD supply chain
ASN + CIDR + Shodan/Censys
email harvesting + leaked credentials
真实性验证技能。分析用户提供的信息、消息、图片或内容，判断其真实性和可信度，识别虚假信息、AI生成内容或伪造内容。
Container escape is a critical attack technique where an adversary breaks out of container isolation to access the host system or other containers. Detection involves monitoring for escape indicators
Pod Security Standards (PSS) define three levels of security policies -- Privileged, Baseline, and Restricted -- enforced by the Pod Security Admission (PSA) controller built into Kubernetes 1.25+. PS
Implements eBPF-based security monitoring using Cilium Tetragon for real-time process execution tracking, network connection observability, file access auditing, and runtime enforcement. Covers TracingPolicy CRD authoring with kprobe/tracepoint hooks, in-kernel filtering via matchArgs/matchBinaries selectors, JSON event export, and integration with SIEM pipelines. Use when building kernel-level runtime security observability for Linux hosts or Kubernetes clusters.
AI Agent Governance Toolkit — Policy enforcement, zero-trust identity, execution sandboxing, and reliability engineering for autonomous AI agents. Covers 10/10 OWASP Agentic Top 10.
Open-source sandboxed agent harness for teams. Giving every employee a secured personal agent.
AI agent security scanner. Detect vulnerabilities in agent configurations, MCP servers, and tool permissions. Available as CLI, GitHub Action, ECC plugin, and GitHub App integration. 🛡️
A secure* runtime for autonomous AI agents. Policy from plain-English constitutions. (*https://ironcurtain.dev)
Learn how to use the absolute-audit Claude skill. Complete guide with installation instructions and examples.
Learn how to use the token-scam-analysis Claude skill. Complete guide with installation instructions and examples.
Learn how to use the absolute-upgrade Claude skill. Complete guide with installation instructions and examples.