{"id":41479626,"url":"https://github.com/intertwine/security-verifiers","last_synced_at":"2026-01-23T17:15:40.672Z","repository":{"id":314359802,"uuid":"1047578740","full_name":"intertwine/security-verifiers","owner":"intertwine","description":"Reinforcement Learning Verifiers for Cybersecurity","archived":false,"fork":false,"pushed_at":"2025-11-21T02:50:10.000Z","size":1385,"stargazers_count":1,"open_issues_count":1,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-11-21T04:14:16.577Z","etag":null,"topics":["cybersecurity","llms","reinforcement-learning"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/intertwine.png","metadata":{"files":{"readme":"README-PRIME.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null}},"created_at":"2025-08-30T18:23:29.000Z","updated_at":"2025-11-21T02:16:07.000Z","dependencies_parsed_at":"2025-10-19T04:12:20.630Z","dependency_job_id":"e27caf28-664a-43d4-aca9-9baee41edf5e","html_url":"https://github.com/intertwine/security-verifiers","commit_stats":null,"previous_names":["intertwine/security-verifiers"],"tags_count":9,"template":false,"template_full_name":null,"purl":"pkg:github/intertwine/security-verifiers","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/intertwine%2Fsecurity-verifiers","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/intertwine%2Fsecurity-verifiers/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/intertwine%2Fsecurity-verifiers/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/intertwine%2Fsecurity-verifiers/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/intertwine","download_url":"https://codeload.github.com/intertwine/security-verifiers/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/intertwine%2Fsecurity-verifiers/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28696521,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-23T15:57:05.722Z","status":"ssl_error","status_checked_at":"2026-01-23T15:56:27.656Z","response_time":59,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["cybersecurity","llms","reinforcement-learning"],"created_at":"2026-01-23T17:15:39.949Z","updated_at":"2026-01-23T17:15:40.666Z","avatar_url":"https://github.com/intertwine.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Prime Intellect Environments Hub Compatibility\n\nThis document assesses the compatibility of the Open Security Verifiers repository with Prime Intellect's Environments Hub and `vf-eval` style evaluations.\n\n## Executive Summary\n\n**Compatibility Status: ✅ HIGH COMPATIBILITY**\n\nThe Open Security Verifiers environments are **already highly compatible** with Prime Intellect's Environments Hub. The repository follows the correct package structure, entry point conventions, and `load_environment()` patterns. However, some adjustments are recommended for optimal integration with `vf-eval` and Hub-based evaluations.\n\n## Current Implementation Analysis\n\n### ✅ What's Already Compatible\n\n1. **Package Structure** (`pyproject.toml`)\n   - All six environments are properly packaged with `pyproject.toml` files\n   - Dependencies are correctly declared\n   - Build system uses hatchling (compatible with pip/uv)\n   - Example from `sv-env-network-logs/pyproject.toml`:\n     ```toml\n     [project.entry-points.\"verifiers.environments\"]\n     sv-env-network-logs = \"sv_env_network_logs:load_environment\"\n     ```\n\n2. **Entry Points Convention**\n   - All environments correctly expose `load_environment()` via `[project.entry-points.\"verifiers.environments\"]`\n   - Entry point naming follows Hub conventions: `owner/env-name` or `env-name`\n   - This matches Prime Intellect's requirements for environment discovery\n\n3. **Environment Interface**\n   - All `load_environment()` functions return proper `verifiers` environment objects:\n     - E1: `vf.SingleTurnEnv`\n     - E2: `vf.ToolEnv`\n     - E3-E6: Various environment types\n   - Functions accept standard parameters (dataset_name, max_examples, logger, etc.)\n   - Parsers, rubrics, and system prompts are correctly implemented\n\n4. **Deployment Infrastructure**\n   - `Makefile` already includes `deploy` target using `prime env push`\n   - Example: `make deploy E=network-logs`\n   - Uses Prime CLI: `prime login` and `prime env push -v PUBLIC`\n\n5. **vf-eval Configuration**\n   - Repository includes `configs/endpoints.py` with endpoint configurations\n   - README files contain `vf-eval` usage examples\n   - Example from `sv-env-network-logs/README-DEV.md`:\n     ```bash\n     vf-eval sv-env-network-logs --model gpt-5-mini --num-examples 3\n     ```\n\n### ⚠️ Compatibility Gaps \u0026 Recommendations\n\n1. **Dataset Loading Strategy**\n   - **Current:** Environments load datasets from local JSONL files\n     - E1: `environments/sv-env-network-logs/data/iot23-train-dev-test-v1.jsonl`\n     - E2: `environments/sv-env-config-verification/data/k8s-labeled-v1.jsonl`\n   - **Issue:** When deployed to Hub, these local files won't be accessible\n   - **Solution:** Implement multi-tiered dataset loading strategy (see Implementation Plan)\n\n2. **vf-eval Testing**\n   - **Current:** No automated tests for `vf-eval` compatibility\n   - **Issue:** Can't verify Hub evaluations work correctly\n   - **Solution:** Add integration tests for `vf-eval` command\n\n3. **Evaluation Script Duplication**\n   - **Current:** Custom evaluation scripts in `scripts/eval_*.py`\n   - **Issue:** Parallel evaluation systems (custom scripts + vf-eval)\n   - **Observation:** Custom scripts provide additional features:\n     - Model routing (OpenAI + OpenRouter)\n     - Early stopping on consecutive errors\n     - Detailed artifact logging (metadata.json, results.jsonl)\n   - **Solution:** Document both workflows; use custom scripts for research, vf-eval for Hub\n\n4. **Hub Deployment Validation**\n   - **Current:** Deployment target exists but not documented/tested\n   - **Issue:** Unknown if current environments can be deployed/installed via Hub\n   - **Solution:** Add deployment validation workflow\n\n## Prime Intellect Requirements (Reference)\n\nBased on [Prime Intellect documentation](https://docs.primeintellect.ai/) and the [Verifiers library](https://github.com/PrimeIntellect-ai/verifiers):\n\n### Environment Structure Requirements\n\n1. **Installable Python Package**\n   - Must have `pyproject.toml` with dependencies ✅\n   - Must be buildable as a wheel ✅\n   - Must be pip/uv installable ✅\n\n2. **Entry Point Registration**\n   - Must expose `load_environment()` via entry points ✅\n   - Format: `[project.entry-points.\"verifiers.environments\"]` ✅\n   - Entry point name becomes the environment identifier ✅\n\n3. **load_environment() Function**\n   - Must return a `verifiers.Environment` subclass ✅\n   - Should accept standard parameters (dataset_name, max_examples, etc.) ✅\n   - Should support synthetic/test datasets for quick testing ✅\n\n4. **Hub Deployment**\n   - Build wheel: `python -m build --wheel` ✅\n   - Authenticate: `prime login` ✅\n   - Deploy: `prime env push -v PUBLIC` ✅\n   - Install: `prime env install owner/env-name` (untested)\n\n5. **vf-eval Compatibility**\n   - Must be loadable via `vf.load_environment(\"env-name\")` ✅\n   - Must work with `vf-eval env-name --model MODEL --num-examples N` (partially tested)\n   - Should handle API keys via environment variables ✅\n\n### Current Compliance\n\n| Requirement | Status | Notes |\n|-------------|--------|-------|\n| Package structure | ✅ Complete | All environments properly packaged |\n| Entry points | ✅ Complete | Correct convention followed |\n| load_environment() | ✅ Complete | Returns proper vf.Environment objects |\n| Build system | ✅ Complete | Wheels build successfully |\n| Deployment target | ⚠️ Partial | Makefile target exists but untested |\n| vf-eval examples | ⚠️ Partial | Documented but not CI-tested |\n| Hub dataset access | ⚠️ Needs work | Local file dependencies |\n| vf-eval integration tests | ❌ Missing | No automated validation |\n\n## Implementation Plan: Full Hub Compatibility\n\n### Phase 1: Dataset Strategy (Priority: High)\n\n**Goal:** Enable environments to work both locally and when deployed to Hub\n\n**Approach:** Implement multi-tiered dataset loading\n\n```python\ndef load_environment(\n    dataset_name: str = \"iot23-train-dev-test-v1.jsonl\",\n    dataset_source: str = \"auto\",  # auto | local | hub | synthetic\n    max_examples: int = 1000,\n    logger: RolloutLogger | None = None,\n) -\u003e vf.SingleTurnEnv:\n    \"\"\"Load environment with flexible dataset sources.\n\n    Dataset loading priority (when source=\"auto\"):\n    1. Local JSONL files (environments/sv-env-*/data/*.jsonl)\n    2. HuggingFace Hub (with HF_TOKEN)\n    3. Synthetic fixtures (for testing)\n    \"\"\"\n    if dataset_source == \"auto\":\n        # Try local first\n        if _local_dataset_exists(dataset_name):\n            dataset = _load_local_dataset(dataset_name, max_examples)\n        # Try Hub second\n        elif _hub_credentials_available():\n            dataset = _load_hub_dataset(dataset_name, max_examples)\n        # Fall back to synthetic\n        else:\n            dataset = _create_synthetic_dataset(max_examples)\n    elif dataset_source == \"local\":\n        dataset = _load_local_dataset(dataset_name, max_examples)\n    elif dataset_source == \"hub\":\n        dataset = _load_hub_dataset(dataset_name, max_examples)\n    elif dataset_source == \"synthetic\":\n        dataset = _create_synthetic_dataset(max_examples)\n\n    # Rest of environment setup...\n```\n\n**Implementation Steps:**\n\n1. Add `dataset_source` parameter to all `load_environment()` functions\n2. Implement HuggingFace Hub dataset loading helper:\n   ```python\n   def _load_hub_dataset(dataset_name: str, max_examples: int):\n       \"\"\"Load dataset from HuggingFace Hub.\"\"\"\n       from datasets import load_dataset\n\n       # Map local names to Hub repos\n       HUB_DATASETS = {\n           \"iot23-train-dev-test-v1.jsonl\": \"intertwine-ai/security-verifiers-e1\",\n           \"k8s-labeled-v1.jsonl\": \"intertwine-ai/security-verifiers-e2\",\n           # ... etc\n       }\n\n       hub_repo = HUB_DATASETS.get(dataset_name)\n       if not hub_repo:\n           raise ValueError(f\"Unknown dataset: {dataset_name}\")\n\n       dataset = load_dataset(hub_repo, split=\"train\", use_auth_token=True)\n       return dataset.select(range(min(len(dataset), max_examples)))\n   ```\n3. Update environment READMEs to document dataset source options\n4. Add integration tests for each dataset source\n\n**Benefits:**\n- Environments work locally (current workflow)\n- Environments work when deployed to Hub (requires HF_TOKEN)\n- Synthetic datasets enable quick testing without data dependencies\n\n### Phase 2: vf-eval Integration Tests (Priority: Medium)\n\n**Goal:** Validate that environments work correctly with `vf-eval` command\n\n**Implementation Steps:**\n\n1. Create `tests/integration/test_vfeval.py`:\n   ```python\n   import subprocess\n   import os\n   import pytest\n\n   @pytest.mark.integration\n   def test_vfeval_e1_synthetic():\n       \"\"\"Test vf-eval with E1 using synthetic dataset.\"\"\"\n       result = subprocess.run([\n           \"vf-eval\", \"sv-env-network-logs\",\n           \"--model\", \"gpt-5-mini\",\n           \"--num-examples\", \"3\",\n           \"--dataset\", \"synthetic\"\n       ], env={**os.environ, \"DATASET_SOURCE\": \"synthetic\"})\n       assert result.returncode == 0\n\n   @pytest.mark.integration\n   def test_vfeval_e2_fixtures():\n       \"\"\"Test vf-eval with E2 using builtin fixtures.\"\"\"\n       result = subprocess.run([\n           \"vf-eval\", \"sv-env-config-verification\",\n           \"--model\", \"gpt-5-mini\",\n           \"--num-examples\", \"2\",\n           \"--dataset\", \"builtin\"\n       ])\n       assert result.returncode == 0\n   ```\n\n2. Add integration test CI workflow:\n   ```yaml\n   # .github/workflows/integration.yml\n   name: Integration Tests\n   on: [push, pull_request]\n   jobs:\n     vfeval:\n       runs-on: ubuntu-latest\n       steps:\n         - uses: actions/checkout@v3\n         - uses: actions/setup-python@v4\n         - run: make setup\n         - run: source .venv/bin/activate\n         - run: pytest tests/integration/ -v -m integration\n   ```\n\n3. Add Makefile target:\n   ```makefile\n   test-integration: venv\n       @$(ECHO) \"$(YELLOW)Running integration tests...$(NC)\"\n       @$(ACTIVATE) \u0026\u0026 pytest tests/integration/ -v -m integration\n   ```\n\n**Benefits:**\n- Automated validation of vf-eval compatibility\n- Catch regressions before Hub deployment\n- Documentation that vf-eval works correctly\n\n### Phase 3: Hub Deployment Validation (Priority: Medium)\n\n**Goal:** Verify environments can be deployed and installed via Hub\n\n**Implementation Steps:**\n\n1. Create deployment validation script (`scripts/validate_hub_deployment.sh`):\n   ```bash\n   #!/bin/bash\n   set -e\n\n   ENV_NAME=$1\n\n   echo \"Building wheel for $ENV_NAME...\"\n   cd environments/sv-env-$ENV_NAME\n   python -m build --wheel\n\n   echo \"Checking wheel contents...\"\n   unzip -l dist/*.whl\n\n   echo \"Testing local installation...\"\n   uv pip install dist/*.whl --force-reinstall\n\n   echo \"Testing load_environment...\"\n   python -c \"import verifiers as vf; env = vf.load_environment('sv-env-$ENV_NAME', dataset_source='synthetic'); print(f'✓ Loaded {env.name}')\"\n\n   echo \"✓ $ENV_NAME validation passed\"\n   ```\n\n2. Add to Makefile:\n   ```makefile\n   validate-env: venv\n       @if [ -z \"$(E)\" ]; then \\\n           echo \"Error: Specify E=env-name\"; exit 1; \\\n       fi\n       @./scripts/validate_hub_deployment.sh $(E)\n   ```\n\n3. Document manual Hub deployment testing:\n   ```markdown\n   ## Testing Hub Deployment\n\n   1. Build and push to Hub (requires prime CLI):\n      ```bash\n      make deploy E=network-logs\n      ```\n\n   2. Install from Hub (in clean environment):\n      ```bash\n      prime env install intertwine/sv-env-network-logs\n      ```\n\n   3. Test with vf-eval:\n      ```bash\n      vf-eval intertwine/sv-env-network-logs --model gpt-5-mini --num-examples 3\n      ```\n   ```\n\n**Benefits:**\n- Confidence that environments deploy successfully\n- Catch packaging issues before Hub upload\n- Documented deployment workflow\n\n### Phase 4: Documentation \u0026 Examples (Priority: High)\n\n**Goal:** Comprehensive documentation for Hub usage\n\n**Implementation Steps:**\n\n1. Create `docs/hub-deployment.md`:\n   - Prerequisites (prime CLI, API keys)\n   - Building wheels\n   - Deploying to Hub\n   - Installing from Hub\n   - Using vf-eval\n   - Troubleshooting common issues\n\n2. Update environment READMEs with Hub examples:\n   ```markdown\n   ## Using from Prime Intellect Hub\n\n   ### Install\n   ```bash\n   prime env install intertwine/sv-env-network-logs\n   ```\n\n   ### Evaluate\n   ```bash\n   # Set API keys\n   export OPENAI_API_KEY=your-key-here\n\n   # Run evaluation\n   vf-eval intertwine/sv-env-network-logs \\\n     --model gpt-5-mini \\\n     --num-examples 10\n   ```\n\n   ### Load in Python\n   ```python\n   import verifiers as vf\n\n   env = vf.load_environment(\n       \"intertwine/sv-env-network-logs\",\n       dataset_source=\"hub\",  # Use Hub datasets\n       max_examples=100\n   )\n   ```\n   ```\n\n3. Add Hub deployment checklist to `CONTRIBUTING.md`:\n   ```markdown\n   ## Hub Deployment Checklist\n\n   - [ ] All tests passing\n   - [ ] Wheel builds successfully\n   - [ ] Local installation works\n   - [ ] vf-eval works with synthetic dataset\n   - [ ] Environment loads correctly\n   - [ ] API keys documented\n   - [ ] README updated with Hub examples\n   - [ ] Version bumped in pyproject.toml\n   ```\n\n4. Create Hub-specific Makefile targets:\n   ```makefile\n   # Validate environment is Hub-ready\n   hub-validate: venv\n       @$(MAKE) validate-env E=$(E)\n       @$(MAKE) test-env E=$(E)\n\n   # Deploy with validation\n   hub-deploy: venv\n       @$(MAKE) hub-validate E=$(E)\n       @$(MAKE) deploy E=$(E)\n   ```\n\n**Benefits:**\n- Clear path for contributors\n- Reduced deployment errors\n- Better user experience\n\n## Comparison: Custom Scripts vs. vf-eval\n\n### Current Custom Scripts (`scripts/eval_*.py`)\n\n**Advantages:**\n- Multi-model routing (OpenAI + OpenRouter APIs)\n- Automatic model name resolution via `model_router.py`\n- Early stopping on consecutive errors (prevents wasted API costs)\n- Structured artifact logging:\n  - `outputs/evals/sv-env-{name}--{model}/{run_id}/metadata.json`\n  - `outputs/evals/sv-env-{name}--{model}/{run_id}/results.jsonl`\n- Dataset selection with validation\n- Git commit tracking in metadata\n- Reproducible evaluations with complete provenance\n\n**Use Cases:**\n- Research evaluations\n- Benchmarking multiple models\n- Cost-sensitive experiments\n- Local development\n\n### vf-eval Command\n\n**Advantages:**\n- Standard Prime Intellect interface\n- Works with Hub-deployed environments\n- Simpler command line interface\n- Integration with Prime Intellect infrastructure\n- Compatible with RL training workflows\n\n**Use Cases:**\n- Hub-based evaluations\n- Quick environment testing\n- Standard benchmarks\n- Integration with Prime RL training\n\n### Recommendation: Support Both\n\n**Strategy:**\n- **Custom scripts:** Primary research evaluation workflow\n- **vf-eval:** Hub compatibility and quick testing\n\n**Implementation:**\n- Keep custom scripts for advanced features\n- Ensure environments work with both approaches\n- Document when to use each\n- Add integration tests for vf-eval\n\n## Testing Roadmap\n\n### Phase 1: Basic Compatibility\n- [ ] Verify entry points are discoverable\n- [ ] Test `vf.load_environment(\"env-name\")` for all environments\n- [ ] Validate synthetic datasets work with vf-eval\n\n### Phase 2: Local vf-eval\n- [ ] Test `vf-eval sv-env-network-logs --model gpt-5-mini --num-examples 3`\n- [ ] Test `vf-eval sv-env-config-verification --model gpt-5-mini --num-examples 2`\n- [ ] Validate API key handling\n- [ ] Test with multiple models\n\n### Phase 3: Hub Deployment\n- [ ] Build and validate all environment wheels\n- [ ] Deploy test environment to Hub\n- [ ] Install from Hub in clean environment\n- [ ] Run vf-eval with Hub-installed environment\n\n### Phase 4: Dataset Integration\n- [ ] Implement Hub dataset loading\n- [ ] Test fallback strategy (local → hub → synthetic)\n- [ ] Validate with production datasets\n- [ ] Test without local data files\n\n### Phase 5: Documentation\n- [ ] Update all environment READMEs\n- [ ] Create Hub deployment guide\n- [ ] Add troubleshooting documentation\n- [ ] Create video walkthrough (optional)\n\n## Quick Start: Testing Hub Compatibility\n\n### Option 1: Test vf-eval Locally (No Hub Required)\n\n```bash\n# Setup\nmake setup\nsource .venv/bin/activate\n\n# Test E1 with synthetic dataset\nvf-eval sv-env-network-logs --model gpt-5-mini --num-examples 3\n\n# Test E2 with builtin fixtures\nvf-eval sv-env-config-verification --model gpt-5-mini --num-examples 2\n```\n\n### Option 2: Test Full Hub Workflow (Requires Prime CLI)\n\n```bash\n# Setup\nuv tool install prime\nprime login\n\n# Build and deploy\nmake deploy E=network-logs\n\n# Install in new environment\ncd /tmp\npython -m venv test_env\nsource test_env/bin/activate\nprime env install intertwine/sv-env-network-logs\n\n# Test\nexport OPENAI_API_KEY=your-key\nvf-eval intertwine/sv-env-network-logs --model gpt-5-mini --num-examples 3\n```\n\n## Recommended Next Steps\n\n### Immediate (Week 1)\n1. ✅ **Document current compatibility status** (this document)\n2. Test basic vf-eval locally with synthetic datasets\n3. Validate entry points work correctly\n4. Run integration tests manually\n\n### Short-term (Weeks 2-4)\n1. Implement multi-tiered dataset loading (Phase 1)\n2. Add vf-eval integration tests (Phase 2)\n3. Update environment READMEs with Hub examples\n4. Test deployment to Hub with one environment\n\n### Medium-term (Weeks 5-8)\n1. Deploy all production-ready environments (E1, E2) to Hub\n2. Create comprehensive Hub deployment guide\n3. Add automated Hub deployment validation\n4. Document troubleshooting procedures\n\n### Long-term (Ongoing)\n1. Monitor Hub deployments for issues\n2. Gather user feedback on vf-eval experience\n3. Optimize dataset loading performance\n4. Integrate with Prime RL training workflows\n\n## Conclusion\n\nThe Open Security Verifiers repository is **already highly compatible** with Prime Intellect's Environments Hub. The package structure, entry points, and environment interfaces follow all required conventions.\n\n**Key Strengths:**\n- ✅ Correct package structure and entry points\n- ✅ Standard `load_environment()` interface\n- ✅ Build and deployment infrastructure in place\n- ✅ vf-eval examples documented\n\n**Minimal Work Required:**\n- ⚠️ Implement multi-tiered dataset loading\n- ⚠️ Add vf-eval integration tests\n- ⚠️ Validate Hub deployment workflow\n\nWith the implementation plan outlined above, full Hub compatibility can be achieved in **2-4 weeks** of focused development.\n\n## References\n\n- [Prime Intellect Verifiers Library](https://github.com/PrimeIntellect-ai/verifiers)\n- [Prime Intellect Environments Hub](https://app.primeintellect.ai/dashboard/environments)\n- [Prime Intellect Documentation](https://docs.primeintellect.ai/)\n- [Prime CLI](https://github.com/PrimeIntellect-ai/prime-cli)\n- [Verifiers Documentation](https://verifiers.readthedocs.io/)\n- [Environments Hub Blog Post](https://www.primeintellect.ai/blog/environments)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fintertwine%2Fsecurity-verifiers","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fintertwine%2Fsecurity-verifiers","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fintertwine%2Fsecurity-verifiers/lists"}