iiRecord
Agentic AI Atlas · Gemini CLI Knowledge Fabric — GEMINI.md
knowledge-fabric-impl:gemini-cli.knowledge@currenta5c.ai
II.
KnowledgeFabricImpl JSON

knowledge-fabric-impl:gemini-cli.knowledge@current

Structured · live

Gemini CLI Knowledge Fabric — GEMINI.md json

Inspect the normalized record payload exactly as the atlas UI reads it.

File · agent-stack/knowledge-fabric-impls/gemini-cli-knowledge-current.yamlCluster · agent-stack
Record JSON
{
  "id": "knowledge-fabric-impl:gemini-cli.knowledge@current",
  "_kind": "KnowledgeFabricImpl",
  "_file": "agent-stack/knowledge-fabric-impls/gemini-cli-knowledge-current.yaml",
  "_cluster": "agent-stack",
  "attributes": {
    "displayName": "Gemini CLI Knowledge Fabric — GEMINI.md",
    "agentVersionId": "agentVersion:gemini:ge-0-0-0",
    "description": "Gemini CLI's knowledge fabric is built on GEMINI.md files that provide\npersistent project-level and user-level instructions. Like Claude Code's\nCLAUDE.md, these files are read at session bootstrap and injected into\nthe system prompt. Gemini CLI supports nested GEMINI.md files in\nsubdirectories for scoped context. The knowledge persists on the\nfilesystem and is retrieved via file-hierarchy traversal.\n",
    "knowledgeFileFormats": [
      "GEMINI.md"
    ],
    "knowledgeScopes": [
      "project",
      "user"
    ],
    "autoExtractionSupport": false,
    "retrievalStrategy": "file-hierarchy",
    "knowledgePersistence": "filesystem",
    "notes": "Gemini CLI follows the AGENTS.md / CLAUDE.md pattern with its own\nGEMINI.md files. Knowledge retrieval is file-hierarchy based with\nproject and user scopes. No auto-memory extraction or vector-based\nretrieval has been documented.\n"
  },
  "outgoingEdges": [
    {
      "from": "knowledge-fabric-impl:gemini-cli.knowledge@current",
      "to": "layer:12-knowledge-fabric",
      "kind": "realizes",
      "attributes": {}
    },
    {
      "from": "knowledge-fabric-impl:gemini-cli.knowledge@current",
      "to": "agentVersion:gemini:ge-0-0-0",
      "kind": "composes",
      "attributes": {
        "role": "knowledge-fabric"
      }
    }
  ],
  "incomingEdges": [
    {
      "from": "agentVersion:gemini:ge-0-0-0",
      "to": "knowledge-fabric-impl:gemini-cli.knowledge@current",
      "kind": "composed_of",
      "attributes": {
        "role": "knowledge-fabric"
      }
    }
  ]
}