{"data":{"slug":"humanbound-humanbound","name":"humanbound","tagline":"Adversarial Testing Engine and SDK for AI Agents","github_url":"https://github.com/humanbound/humanbound","owner":"humanbound","repo":"humanbound","owner_avatar_url":"https://avatars.githubusercontent.com/u/125462661?v=4","primary_language":"Python","stars":144,"forks":16,"topics":["adversarial-testing","agentic-ai","ai-agents","ai-red-teaming","ai-safety","ai-security","aisecops","cli","cybersecurity","guardrails","humanbound","llm","llm-security","multimodal-ai","multimodal-security","owasp","pentesting","prompt-injection","security-testing","security-tools"],"archived":false,"github_pushed_at":"2026-09-09T17:44:54+00:00","maintenance_label":"Very active","stars_delta_30d":26,"url":"https://www.graphcanon.com/tools/humanbound-humanbound","markdown_url":"https://www.graphcanon.com/tools/humanbound-humanbound.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/humanbound-humanbound","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=humanbound-humanbound","description":"Open-source adversarial testing engine, SDK, and CLI for AI agents. Runs locally or against the Humanbound Platform.","homepage_url":"https://docs.humanbound.ai/","license":"Other","open_issues":12,"watchers":0,"ai_summary":"An open-source adversarial testing engine that provides an SDK and CLI to test the robustness of AI agents locally or via the Humanbound Platform.","readme_excerpt":"### Install\n\n```bash\npip install humanbound                       # CLI + SDK, core deps\npip install humanbound[engine]               # + OpenAI / Anthropic / Gemini / Ollama providers\npip install humanbound[firewall]             # + humanbound-firewall runtime\npip install humanbound[engine,firewall]      # everything\n```\n\n---\n\n### Docker\n\nRun the CLI from the official image ([docs](https://docs.humanbound.ai/integrations/docker/)):\n\n```bash\ndocker run --rm -v \"$PWD\":/workspace \\\n  -e HB_PROVIDER=openai -e HB_API_KEY=$OPENAI_API_KEY -e HB_MODEL=gpt-4o-mini \\\n  ghcr.io/humanbound/humanbound:2 test --endpoint ./bot-config.json --wait\n```\n\n---\n\n## License\n\n[Apache-2.0](https://github.com/humanbound/humanbound/blob/main/LICENSE). Free to use in any context — commercial or\nopen-source — with attribution. See [TRADEMARK.md](https://github.com/humanbound/humanbound/blob/main/TRADEMARK.md) for the\ntrademark policy. The code is open; the name is not.\n\nThe sibling project [`humanbound-firewall`](https://github.com/humanbound/humanbound-firewall)\nis also Apache-2.0 — same license, different product.","github_created_at":"2026-02-08T08:47:43+00:00","created_at":"2026-07-15T10:44:54.757958+00:00","updated_at":"2026-09-20T04:26:30.017971+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":"evaluation-observability","name":"Evaluation & Observability","url":"https://www.graphcanon.com/categories/evaluation-observability","markdown_url":"https://www.graphcanon.com/categories/evaluation-observability.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/evaluation-observability"}],"tags":[{"slug":"adversarial-testing","name":"adversarial-testing"},{"slug":"agentic-ai","name":"agentic-ai"},{"slug":"ai-agents","name":"ai-agents"},{"slug":"llm-security","name":"llm-security"},{"slug":"multimodal-ai","name":"multimodal-ai"},{"slug":"security-testing","name":"security-testing"}],"trust":{"provenance":{"is_fork":false,"github_id":1152676785,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-13T06:00:43.691Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":8,"days_since_push":3,"last_release_at":"2026-08-14T13:28:59Z","stars_delta_30d":26,"open_issues_delta_30d":2},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-15T10:44:56.374Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-09-13T06:00:44.185Z"},"has_cli":{"value":true,"source":"pyproject.toml:[project.scripts]","observed_at":"2026-09-13T06:00:44.185Z"},"languages":{"value":["python"],"source":"github.language+pyproject.toml","observed_at":"2026-09-13T06:00:44.185Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-09-13T06:00:44.185Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When you need to test your AI agent's resilience against prompt injection attacks, utilize humanbound’s specialized features tailored for this purpose","For scenarios that require comprehensive multimodal security testing, including text, image and other modalities, humanbound offers detailed adversarial testing capabilities"],"when_not_to_use":["Avoid using humanbound if your project does not involve AI agents or is not concerned about adversarial robustness since the tool's functionality might be overly specific","Do not use humanbound in environments where an open-source solution is restricted, particularly considering its licensing and trademark policies"],"source":"enrich:decision_facts","observed_at":"2026-07-16T18:41:36.998Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"humanbound is an adversarial testing engine and SDK in Python designed specifically for evaluating the robustness of AI agents against various security threats."}]}}