Update to Elephant-Goldfish Model

{
  "description": "Compare checked-in documents with the current source code.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 304
    },
    {
      "relative_path": "assets/icon.png",
      "size_in_bytes": 4333
    },
    {
      "relative_path": "assets/icon.svg",
      "size_in_bytes": 281
    }
  ],
  "name": "egm-drift-report",
  "skill_md_contents": "---\nname: egm-drift-report\nlicense: Apache-2.0\ndisable-model-invocation: true\ndescription: Compare checked-in documents with the current source code.\n---\n\n# Detect Drift In Designs and Code\n\nWhen running in Codex or ChatGPT Work, first read [Codex integration](../../docs/codex.md).\nIn Claude Code or Cowork, read `${CLAUDE_PLUGIN_ROOT}/docs/claude.md` first.\n\nWhen the user asks to detect drift across all designs or a set of READMEs, the\ngoal is to identify all instances where the codebase has diverged or failed to\nimplement the specifications described by those designs or READMEs.\n\n1.  **Check Input**: Verify a directory containing design documents or READMEs\n    is provided. If not provided, ask the user to specify the target directory.\n2.  **List Documents**: List all files in the design directory.\n3.  **Analyze**: For EACH design document:\n\n    *   **Contextualize**: Read the design document. Identify the intent,\n        proposed changes, and specific filenames/symbols mentioned.\n    *   **Locate Source**: Find the corresponding source code files in the\n        source tree. If paths aren't explicit, use code search or grep to\n        locate the relevant classes, functions, or API endpoints mentioned in\n        the doc. Prompt the user to add missing links to the document to make\n        future detection cheaper.\n    *   **Verify**: Read the current state of those source files.\n    *   **Detect Drift**: Compare the design vs. reality. Look specifically for:\n        *   Missing Features: Logic described in the doc but absent in the code.\n        *   Incomplete Models: Dataclasses/Classes/Structs/Protos missing fields\n            specified in the design.\n        *   Logic Shortcuts: Hardcoded values where dynamic logic was planned.\n        *   Architectural Divergence: Implementation using different\n            patterns/libraries than proposed.\n    *   **Filter**: Report ONLY the differences. Do not describe what is working\n        or what matches. If the code matches the design perfectly, output \"No\n        drift detected.\"\n\n4.  **Report**: Produce a single report listing all discrepancies found.\n\n    *   Append the analysis for each file to `{YYYY-MM-DD}_DRIFT_REPORT.md`\n        using your file-writing tool.\n    *   Strict Format:\n\n        ```markdown\n        ## <DESIGN_DOC_NAME>\n\n        <Bulleted list of specific divergences found>\n        <Or \"No drift detected\">\n        ```\n\n5.  **Guidance**\n\n    -   Be rigorous. If a design doc mentions a specific object, field or file\n        name and the code doesn't have it, that is a reportable drift. Double\n        check for coverage. If there is some part of the logic that is not\n        referenced in a design or README, this is a reportable drift.\n    -   Read files liberally to ensure you are looking at the current state\n        of the code, not relying on memory.\n    -   Process the files sequentially to ensure the tool output buffer doesn't\n        overflow.\n"
}