{"data":{"slug":"0xsteph-pentest-ai","name":"pentest-ai","tagline":"Offensive-security MCP server with security tools and agents","github_url":"https://github.com/0xSteph/pentest-ai","owner":"0xSteph","repo":"pentest-ai","owner_avatar_url":"https://avatars.githubusercontent.com/u/245668117?v=4","primary_language":"Python","stars":1685,"forks":314,"topics":["ai-penetration-testing","ai-pentesting","ai-security","appsec","bug-bounty","claude","ctf","cybersecurity","devsecops","ethical-hacking","hacking-tools","mcp","model-context-protocol","offensive-security","penetration-testing","pentesting","python","red-team","security","vulnerability-scanner"],"archived":false,"github_pushed_at":"2026-09-13T07:44:24+00:00","maintenance_label":"Very active","stars_delta_30d":244,"url":"https://www.graphcanon.com/tools/0xsteph-pentest-ai","markdown_url":"https://www.graphcanon.com/tools/0xsteph-pentest-ai.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/0xsteph-pentest-ai","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=0xsteph-pentest-ai","description":"Open-source AI pentester that proves every finding. Machine oracles re-run each exploit; verified bugs ship a proof capsule you can replay yourself.","homepage_url":"https://pentestai.xyz","license":"MIT","open_issues":2,"watchers":13,"ai_summary":"Pentest-ai offers an offensive-security platform that includes various security tools for bug bounty, CTF, OSINT tasks, exploit chaining, and more.","readme_excerpt":"[Website](https://pentestai.xyz/) · [Quick start](#quick-start) · [Verification](#how-verification-works) · [Documentation](https://github.com/0xSteph/pentest-ai/tree/main/docs) · [Issues](https://github.com/0xSteph/pentest-ai/issues)\n\nPentest-AI connects an AI client or model to a penetration-testing workflow. It investigates applications, records findings, and uses **machine oracles to check whether an exploit can be reproduced**. Verified findings carry evidence you can replay.\n\nUse it through **MCP**, from the **CLI**, or in **CI**. The model helps drive the investigation; an oracle determines whether a finding earns a verified verdict.\n\n> Test only systems you own or have explicit authorization to assess. Read the [acceptable-use policy](https://pentestai.xyz/aup) and [terms](https://pentestai.xyz/terms) before running an engagement.\n\n## Quick start\n\n### Recommended: Codex or Claude Code\n\nUse Python **3.10–3.14** and install [pipx](https://pipx.pypa.io/stable/installation/)\nfirst. Run `pipx ensurepath` and reopen your terminal if needed. pipx isolates\nPentest-AI's dependencies and makes its commands available across folders.\n\n```sh\npipx install ptai\nptai setup --mcp\n```\n\nInstall your AI client's CLI first. Setup asks before connecting **all detected\nsupported clients**, including both Codex and Claude Code when installed. It\npreserves existing registrations. Restart your client afterward.\n\nThe wizard uses **user-wide settings**. It does not install clients, sign you in,\ndownload models or configure Ollama. Your AI client supplies the model; no separate\nmodel API key is required by the MCP server. Use `ptai setup --mcp --dry-run` to\npreview changes. See [single-client and per-repository setup](https://pentestai.xyz/docs/ai-clients/)\nfor Codex `.codex/config.toml` and Claude Code `.mcp.json` instructions.\n\nFor upgrades, use `pipx upgrade ptai`. In an existing activated virtual environment,\n`python -m pip install --upgrade ptai` is also supported; keep that environment\navailable after registering it with your AI client.\n\n### Alternative: Ollama and local models\n\nInstall Pentest-AI as above, but skip MCP setup for this route.\n[Install and start Ollama](https://docs.ollama.com/quickstart), then pull a model\nwith tool-calling support. Use `ollama list` to find its exact installed name.\n\n```bash\nexport PENTEST_AI_LLM_PROVIDER=ollama\nexport PENTEST_AI_MODEL=\"your-installed-model\"\nptai start https://your-authorized-target.example --no-sync\n```\n\nIn PowerShell:\n\n```powershell\n$env:PENTEST_AI_LLM_PROVIDER = \"ollama\"\n$env:PENTEST_AI_MODEL = \"your-installed-model\"\nptai start https://your-authorized-target.example --no-sync\n```\n\nReplace the placeholders before running. A local model needs no cloud API key.\nModel quality and available memory affect results. See the\n[model guide](https://pentestai.xyz/docs/ai-clients/) for endpoint configuration.\n\n### Try the bundled demo\n\nNo API key or external target required.\n\n```bash\npipx install ptai\nptai demo\n```\n\nThe demo tests a bundled vulnerable application, replays a finding from its proof capsule, and checks the hardened routes. Inspect the evidence workflow before configuring a real engagement.\n\n### Other model providers\n\nFor standalone cloud models and optional scanner dependencies, see the [installation guide](https://pentestai.xyz/docs/getting-started/) and [provider guide](docs/llm-providers.md). Keep API keys out of source control.\n\n## How verification works\n\n\n\n1. **Find a candidate.** A probe or scanner identifies behaviour worth investigating.\n2. **Reproduce it and check a control.** A named oracle repeats the exploit and compares it with a control that should not demonstrate the weakness.\n3. **Keep the evidence.** A verified finding records the oracle and its results in a portable proof capsule. Use `ptai replay` to check it again.\n\nFor example, a trusted-header bypass must return privileged content with the attack header and deny access without it. A server that answers every request","github_created_at":"2026-04-04T06:58:30+00:00","created_at":"2026-07-11T11:52:19.372199+00:00","updated_at":"2026-09-30T21:18:58.33106+00:00","categories":[{"slug":"ai-agents","name":"AI Agents","url":"https://www.graphcanon.com/categories/ai-agents","markdown_url":"https://www.graphcanon.com/categories/ai-agents.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/ai-agents"},{"slug":"developer-tools","name":"Developer Tools","url":"https://www.graphcanon.com/categories/developer-tools","markdown_url":"https://www.graphcanon.com/categories/developer-tools.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/developer-tools"}],"tags":[{"slug":"cybersecurity","name":"cybersecurity"},{"slug":"mcp","name":"mcp"},{"slug":"pentesting","name":"pentesting"},{"slug":"security","name":"security"}],"trust":{"provenance":{"is_fork":false,"github_id":1201069034,"owner_type":"User","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-20T03:54:58.246Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":7,"days_since_push":6,"last_release_at":"2026-09-02T03:27:01Z","stars_delta_30d":244,"open_issues_delta_30d":0},"security_summary":{"status":"no_manifest","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T11:52:21.124Z","medium_count":0,"scan_profile":"mcp_manifest","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-09-20T03:54:59.417Z"},"deploy":{"source":"dockerfile:Dockerfile","self_host":true,"observed_at":"2026-09-20T03:54:59.417Z","managed_saas":false},"has_cli":{"value":true,"source":"pyproject.toml:[project.scripts]","observed_at":"2026-09-20T03:54:59.417Z"},"languages":{"value":["python"],"source":"github.language+pyproject.toml","observed_at":"2026-09-20T03:54:59.417Z"},"has_docker":{"value":true,"source":"dockerfile:Dockerfile","observed_at":"2026-09-20T03:54:59.417Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-09-20T03:54:59.417Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When you need a comprehensive suite that includes more than 205 security tools and 17 specialist agents.","For projects requiring exploitation techniques covered by the OWASP Top 10, leveraging the platform's 60 SPA-aware probes."],"when_not_to_use":["If your requirements do not include offensive-security capabilities or exploit chaining.","When you want to avoid MIT-licensed open-source code and prefer proprietary solutions without sharing obligations."],"source":"enrich:decision_facts","observed_at":"2026-07-17T12:34:36.983Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"Pentest-ai is an offensive-security platform featuring over 200 security tools and agents for various security tasks including exploit chaining and osint, licensed under MIT."}]}}