Skip to content

[Bug]: document-extract plugin loads but registers zero capabilities (bundled, 2026.6.1) #91539

Description

@CEKitchens

Bug type

Regression (worked before, now fails)

Beta release blocker

No

Summary

The bundled document-extract plugin loads but registers zero capabilities after upgrading to 2026.6.1, making DOCX and PDF file reading non-functional.

Steps to reproduce

  1. Install/upgrade to OpenClaw 2026.6.1
  2. Run: openclaw plugins inspect document-extract --runtime --json
  3. Observe contracts.documentExtractors declares ["pdf"] but capabilityCount is 0,
    tools is [], and shape is "non-capability"
  4. Plugin is absent from gateway startup log despite showing status "loaded"

Expected behavior

The document-extract plugin registers its PDF/DOCX extraction capability at gateway startup and appears in the gateway startup plugins list, allowing the agent to read local document files.

Actual behavior

Plugin shows status "loaded" in openclaw plugins inspect but registers zero capabilities: capabilityCount: 0, tools: [], shape: "non-capability". Plugin is absent from gateway startup log. Agent cannot read DOCX or PDF files. Began after upgrading from 2026.5.27 to 2026.6.1.

OpenClaw version

2026.6.1

Operating system

Windows 11

Install method

npm global

Model

anthropic/claude-haiku-4-5-20251001

Provider / routing chain

openclaw -> anthropic

Additional provider/model setup details

Logs, screenshots, and evidence

PS C:\WINDOWS\system32> openclaw plugins inspect document-extract --runtime --json
{
  "workspaceDir": "C:\\Users\\Effie Stone\\.openclaw\\workspace",
  "plugin": {
    "id": "document-extract",
    "name": "Document Extraction",
    "description": "Extract text and fallback page images from local document attachments.",
    "version": "2026.6.1",
    "packageName": "@openclaw/document-extract-plugin",
    "format": "openclaw",
    "source": "C:\\Users\\Effie Stone\\AppData\\Roaming\\npm\\node_modules\\openclaw\\dist\\extensions\\document-extract\\index.js",
    "rootDir": "C:\\Users\\Effie Stone\\AppData\\Roaming\\npm\\node_modules\\openclaw\\dist\\extensions\\document-extract",
    "origin": "bundled",
    "enabled": true,
    "compat": [],
    "explicitlyEnabled": true,
    "activated": true,
    "activationSource": "explicit",
    "activationReason": "enabled in config",
    "syntheticAuthRefs": [],
    "status": "loaded",
    "toolNames": [],
    "hookNames": [],
    "channelIds": [],
    "cliBackendIds": [],
    "providerIds": [],
    "embeddingProviderIds": [],
    "speechProviderIds": [],
    "realtimeTranscriptionProviderIds": [],
    "realtimeVoiceProviderIds": [],
    "mediaUnderstandingProviderIds": [],
    "transcriptSourceProviderIds": [],
    "imageGenerationProviderIds": [],
    "videoGenerationProviderIds": [],
    "musicGenerationProviderIds": [],
    "webFetchProviderIds": [],
    "webSearchProviderIds": [],
    "migrationProviderIds": [],
    "contextEngineIds": [],
    "memoryEmbeddingProviderIds": [],
    "agentHarnessIds": [],
    "cliCommands": [],
    "services": [],
    "gatewayDiscoveryServiceIds": [],
    "commands": [],
    "httpRoutes": 0,
    "hookCount": 0,
    "configSchema": true,
    "configJsonSchema": {
      "type": "object",
      "additionalProperties": false,
      "properties": {}
    },
    "contracts": {
      "documentExtractors": [
        "pdf"
      ]
    },
    "imported": true,
    "dependencyStatus": {
      "hasDependencies": false,
      "installed": true,
      "requiredInstalled": true,
      "optionalInstalled": true,
      "missing": [],
      "missingOptional": [],
      "dependencies": [],
      "optionalDependencies": []
    }
  },
  "shape": "non-capability",
  "capabilityMode": "none",
  "capabilityCount": 0,
  "capabilities": [],
  "typedHooks": [],
  "customHooks": [],
  "tools": [],
  "commands": [],
  "cliCommands": [],
  "services": [],
  "gatewayDiscoveryServices": [],
  "gatewayMethods": [],
  "mcpServers": [],
  "lspServers": [],
  "httpRouteCount": 0,
  "bundleCapabilities": [],
  "diagnostics": [],
  "policy": {
    "allowedModels": [],
    "hasAllowedModelsConfig": false
  },
  "usesLegacyBeforeAgentStart": false,
  "compatibility": []
}
PS C:\WINDOWS\system32>

Impact and severity

Affected: any agent using the bundled document-extract plugin on 2026.6.1
Severity: High — blocks all DOCX and PDF file reading
Frequency: Always (100% reproducible)
Consequence: Agent cannot process any document files; workaround requires manually attaching files via Telegram

Additional information

Last known good version: 2026.5.27
First known bad version: 2026.6.1
Plugin origin is "bundled" — cannot be reinstalled independently.
contracts.documentExtractors declares ["pdf"] but capabilityCount is 0 and tools is empty, suggesting capability registration silently fails during gateway startup.

Metadata

Metadata

Assignees

No one assigned

    Labels

    P1High-priority user-facing bug, regression, or broken workflow.bugSomething isn't workingclawsweeper:linked-pr-openClawSweeper found an open linked pull request for this issue.clawsweeper:needs-live-reproClawSweeper needs live local, crabbox, or manual validation to confirm this issue.clawsweeper:no-new-fix-prClawSweeper does not recommend queueing a new automated fix PR for this issue.impact:otherThis issue has meaningful maintainer-visible impact outside the owned taxonomy.issue-rating: 🐚 platinum hermitGood issue quality with a plausible reproduction path needing some confirmation.regressionBehavior that previously worked and now failsstaleMarked as stale due to inactivity

    Type

    No type

    Fields

    Priority

    None yet

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions