toeverything/AFFiNE · error · ToolError

DOC_READ_FAILED

DOC_READ_FAILED

Error message

The persisted Page content could not be read.

What it means

Catch-all in the doc_read tool's execute: the underlying getDoc call threw (storage/decode failure), so the error is logged with the doc id and a structured DOC_READ_FAILED tool error is returned to the model.

Source

Thrown at packages/backend/server/src/plugins/copilot/tools/doc-read.ts:140

      .object({
        doc_id: z
          .string()
          .min(1)
          .max(128)
          .describe('The persisted document to read'),
        max_chars: z.number().int().min(1).max(100_000).optional(),
      })
      .strict(),
    execute: async ({ doc_id, max_chars }) => {
      try {
        const doc = await getDoc(
          doc_id,
          Math.min(max_chars ?? 40_000, 100_000)
        );
        return isToolError(doc) ? doc : { ...doc };
      } catch {
        logger.error(`Failed to read doc ${doc_id}: DOC_READ_FAILED`);
        return toolError(
          'Doc Read Failed',
          'The persisted Page content could not be read.',
          { code: 'DOC_READ_FAILED', retryable: false, locator: { doc_id } }
        );
      }
    },
  });
};

View on GitHub (pinned to b4c8548c09)

Solutions

  1. Retry reading the persisted page content; check storage health.
  2. Verify the document exists in the workspace.
Defensive patterns

Strategy: try-catch

When it happens

Trigger: Returned as DOC_READ_FAILED when getDoc throws unexpectedly while loading persisted page content inside the tool execute wrapper.

Common situations: An internal error prevented loading the page content for the copilot. Retry; if it persists, inspect server logs for DOC_READ_FAILED.

AI-assisted analysis of toeverything/AFFiNE@b4c8548c09 (2026-08-18). Data as JSON: /api/errors/341c00f66f62bfbe. Report an issue: GitHub.