{
  "id": "phi-os-corpus-rhizome",
  "title": "PHI|OS Corpus Rhizome",
  "objective": "Bind canonical PHI|OS corpora to runtime dataset links so interview agents, RAG, and sparse training share one grounded dataset surface.",
  "root": "/rhiz/PHI|OS/corpus",
  "syncScript": "/rhiz/akatsuki/singularity/deploy/scripts/phi-os-corpus-sync.sh",
  "datasets": [
    {
      "id": "phi-os-lab-interviews",
      "kind": "local-jsonl",
      "canonicalPath": "/rhiz/factory/hypergraph_meta_cluster_bundle/services/shared/datasets/phi_os_lab_interviews.jsonl",
      "runtimePath": "/data/datasets/phi-os-lab-interviews",
      "role": "Analytical psychology and orchestration seed corpus tracked in the repository."
    },
    {
      "id": "open-rl",
      "kind": "huggingface-dataset",
      "uri": "https://huggingface.co/datasets/TuringEnterprises/Open-RL",
      "canonicalPath": "/rhiz/PHI|OS/corpus/huggingface/Open-RL",
      "runtimePath": "/rhiz/factory/hypergraph_meta_cluster_bundle/data/datasets/open-rl",
      "syncCommand": "/rhiz/akatsuki/singularity/deploy/scripts/phi-os-corpus-sync.sh sync open-rl",
      "role": "Sensitive RAG and policy-memory substrate."
    },
    {
      "id": "edge-agent-reasoning-websearch-260k",
      "kind": "huggingface-dataset",
      "uri": "https://huggingface.co/datasets/yatin-superintelligence/Edge-Agent-Reasoning-WebSearch-260K",
      "canonicalPath": "/rhiz/PHI|OS/corpus/huggingface/Edge-Agent-Reasoning-WebSearch-260K",
      "runtimePath": "/rhiz/factory/hypergraph_meta_cluster_bundle/data/datasets/edge-agent-reasoning-websearch-260k",
      "syncCommand": "/rhiz/akatsuki/singularity/deploy/scripts/phi-os-corpus-sync.sh sync edge-agent-reasoning-websearch-260k",
      "role": "Agentic reasoning and web-search trajectory corpus for interviewer decomposition, tool-use grounding, and edge-agent evaluation."
    }
  ],
  "constraints": [
    "Keep `/rhiz/PHI|OS/corpus` as the canonical host-side source of truth.",
    "Expose runtime links under `/rhiz/factory/hypergraph_meta_cluster_bundle/data/datasets` for compose-mounted services.",
    "Do not claim local availability until the sync script has run successfully."
  ],
  "bindings": [
    {
      "surface": "Data Rhizome",
      "role": "Curated dataset registry and corpus access"
    },
    {
      "surface": "PHI|OS Control Plane",
      "role": "Corpus management and sync visibility"
    },
    {
      "surface": "Sparse Trainer",
      "role": "Local dataset path for future PHI|OS training jobs"
    }
  ]
}