{"data":{"slug":"santanderai-autoguardrails","name":"autoguardrails","tagline":"Alignment-research scaffold for LLM guardrails involving policy evaluation and content moderation","github_url":"https://github.com/SantanderAI/autoguardrails","owner":"SantanderAI","repo":"autoguardrails","owner_avatar_url":"https://avatars.githubusercontent.com/u/284004589?v=4","primary_language":"Python","stars":129,"forks":36,"topics":["ai","ai-safety","alignment","autoresearch","benchmark","content-moderation","evaluation","guardrails","jailbreak","llm","llm-safety","machine-learning","prompt-injection","python","red-teaming","responsible-ai","santander","security"],"archived":false,"github_pushed_at":"2026-10-01T04:22:30+00:00","maintenance_label":"Active","stars_delta_30d":-1,"url":"https://www.graphcanon.com/tools/santanderai-autoguardrails","markdown_url":"https://www.graphcanon.com/tools/santanderai-autoguardrails.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/santanderai-autoguardrails","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=santanderai-autoguardrails","description":"Alignment-research scaffold (autoresearch-style) for LLM guardrails: search over a single policy.md surface","homepage_url":"https://github.com/SantanderAI","license":"Apache-2.0","open_issues":1,"watchers":5,"ai_summary":"Autoguardrails is scaffolding for alignment research to develop LLM safety mechanisms like guardrails through an autoresearch style workflow. It operates on a policy.md file to evaluate different policies against preset criteria, allowing for iterative adjustment and testing in a controlled environment.","readme_excerpt":"## Quick Start\n\nRun from the repository root.\n\n1. Record a baseline.\n\n```bash\npython -m autoguardrails baseline --reset --repeat 2 --notes \"initial baseline\"\n```\n\n2. Edit only `policy.md`.\n\n3. Score the new candidate.\n\n```bash\npython -m autoguardrails candidate --repeat 2 --notes \"cover jailbreak and obfuscation\"\n```\n\n4. Inspect the current kept result.\n\n```bash\npython -m autoguardrails status\n```\n\n5. Inspect the full log.\n\n```bash\ncat results.tsv\n```\n\nIf a candidate is rejected, the harness restores `policy.md` to the last accepted version automatically.\n\n---\n\n## Requirements\n\n- **Python 3.10+**\n- **No third-party runtime dependencies** — the harness is built entirely on the Python standard library and runs offline by default.\n- Optional, for development only: `ruff`, `black`, `mypy`, `pytest`, `pytest-cov` (see [CONTRIBUTING.md](CONTRIBUTING.md)).\n- Optional, for real-model experiments: access to an OpenAI-compatible chat-completions endpoint (configured via the `AUTOGUARDRAILS_*` environment variables described above).\n\n---\n\n## License\n\nThis project is licensed under the **Apache License 2.0** — see the [LICENSE](LICENSE)\nand [NOTICE](NOTICE) files for details.\n\n```\nCopyright (c) 2026 Santander Group\nSPDX-License-Identifier: Apache-2.0\n```","github_created_at":"2026-06-15T13:10:50+00:00","created_at":"2026-07-15T10:43:17.868929+00:00","updated_at":"2026-10-21T06:00:43.008401+00:00","categories":[{"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"},{"slug":"llm-frameworks","name":"LLM Frameworks","url":"https://www.graphcanon.com/categories/llm-frameworks","markdown_url":"https://www.graphcanon.com/categories/llm-frameworks.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/llm-frameworks"}],"tags":[{"slug":"ai-safety","name":"ai-safety"},{"slug":"alignment","name":"alignment"},{"slug":"autoresearch","name":"autoresearch"},{"slug":"content-moderation","name":"content-moderation"},{"slug":"evaluation","name":"evaluation"},{"slug":"llm-safety","name":"llm-safety"},{"slug":"red-teaming","name":"red-teaming"},{"slug":"responsible-ai","name":"responsible-ai"}],"trust":{"provenance":{"is_fork":false,"github_id":1270177900,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-10-21T06:00:41.489Z","maintenance":{"label":"Active","score":82,"methodology":"github_public_v1","releases_90d":0,"days_since_push":20,"last_release_at":"2026-06-16T20:24:49Z","stars_delta_30d":-1,"open_issues_delta_30d":-1},"security_summary":{"last_scan_at":"2026-10-18T04:00:54.833793+00:00","scan_profile":"deps_dev","status":"no_lockfile","critical_count":0,"high_count":0,"medium_count":0,"low_count":0,"scanner":"deps.dev@v1","scan_id":"541ee31d-91c3-46d1-be51-c97a0c0eb8c0","pending_review_count":0}},"capability_facts":{"mcp":{"method":"github-tree-and-contents@2","reason":"No explicit MCP server declaration was found in the inspected manifests. Other layouts and implementations were not tested.","source":"repo_scan","status":"unknown","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:41.898Z","server_manifest":null},"scan":{"method":"github-tree-and-contents@2","source":"repo_scan","status":"complete","version":2,"coverage":{"scope":"bounded-repository-manifests","tree_complete":true,"selected_files":1,"candidate_files":1},"revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:41.898Z","observations":[{"path":"@revision","bytes":5217,"status":"present","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/commit/1ca0c9b2c0052080f07031247603642074996025","http_status":200,"observed_at":"2026-10-21T06:00:42.139Z"},{"path":"@tree","bytes":15193,"status":"present","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","http_status":200,"observed_at":"2026-10-21T06:00:42.448Z"},{"path":"docker-compose.yml","reason":"not-listed-in-complete-tree","status":"not-found","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:42.448Z"},{"path":"Dockerfile","reason":"not-listed-in-complete-tree","status":"not-found","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:42.448Z"},{"path":"mcp.json","reason":"not-listed-in-complete-tree","status":"not-found","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:42.448Z"},{"path":"package.json","reason":"not-listed-in-complete-tree","status":"not-found","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:42.448Z"},{"path":"pyproject.toml","bytes":2500,"status":"present","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/blob/1ca0c9b2c0052080f07031247603642074996025/pyproject.toml","http_status":200,"observed_at":"2026-10-21T06:00:42.586Z"},{"path":"server.json","reason":"not-listed-in-complete-tree","status":"not-found","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:42.448Z"}]},"has_cli":{"value":true,"method":"repository-declaration@2","source":"pyproject.toml:[project.scripts]","status":"declared","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/blob/1ca0c9b2c0052080f07031247603642074996025/pyproject.toml","observed_at":"2026-10-21T06:00:41.898Z"},"languages":{"value":["python"],"method":"github-tree-and-contents@2","source":"github.language+pyproject.toml","revision":"1ca0c9b2c0052080f07031247603642074996025","source_url":"https://github.com/SantanderAI/autoguardrails/tree/1ca0c9b2c0052080f07031247603642074996025","observed_at":"2026-10-21T06:00:41.898Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-10-21T06:00:41.898Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":{"notes":["Requires Python 3.10 or higher.","No third-party runtimes; it is built completely on the standard Python library."],"min_ram_gb":null,"requires_docker":false},"constraints":{"min_ram_gb":null,"requires_docker":false},"when_to_use":["When you are conducting alignment research that requires systematic iteration on LLM safeguard policies.","If your project operates with an open-source requirement, given its Apache-2.0 license agreement."],"when_not_to_use":["Autoguardrails may not suit needs requiring real-time or dynamic policy adjustments outside its autoresearch workflow.","Avoid using Autoguardrails if you cannot accept offline operation as it is built on the Python standard library and runs without third-party runtime dependencies."],"source":"enrich:decision_facts","observed_at":"2026-07-15T11:11:48.463Z"},"constraint_facets":{"min_ram_gb":null,"requires_docker":false},"decision_summary":[{"label":"Requirements","value":"Requires Python 3.10 or higher.; No third-party runtimes; it is built completely on the standard Python library."},{"label":"Adopt for","value":"Unverified catalog guidance: Autoguardrails is an evaluation and development framework for AI policy creation and review. It enables the iterative adjustment and testing of guardrail policies in alignment research through a controlled workflow."}]}}