fix(files): read offset 0 as the first line instead of rejecting it (#8588)

* fix(files): read offset 0 as the first line instead of rejecting it

* test(files): cover the agent CLI forwarding --offset 0

Author: waleedlatif1

apps/docs/content/docs/cli/files.mdx

@@ -271,7 +271,7 @@ sim files versions read <fileId> <version> [options]
 | Option | Required | Description |
 | --- | --- | --- |
 | `--max-bytes <value>` | No | Optional ceiling on the source bytes fed to the parser, lowering but never raising the server limit. |
-| `--offset <value>` | No | First line to return, 1-based. Absent starts at the first line. |
+| `--offset <value>` | No | First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line. |
 | `--limit <value>` | No | How many lines to return from `offset`. Absent reads to the end. |
 
 </CommandTable>
@@ -497,7 +497,7 @@ sim files read <fileId> [options]
 | Option | Required | Description |
 | --- | --- | --- |
 | `--max-bytes <value>` | No | Optional ceiling on the source bytes fed to the parser, lowering but never raising the server limit. |
-| `--offset <value>` | No | First line to return, 1-based. Absent starts at the first line. |
+| `--offset <value>` | No | First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line. |
 | `--limit <value>` | No | How many lines to return from `offset`. Absent reads to the end. |
 
 </CommandTable>

apps/docs/content/docs/cli/reference.mdx

@@ -1039,7 +1039,7 @@ sim files versions read <fileId> <version> [options]
 | Option | Required | Description |
 | --- | --- | --- |
 | `--max-bytes <value>` | No | Optional ceiling on the source bytes fed to the parser, lowering but never raising the server limit. |
-| `--offset <value>` | No | First line to return, 1-based. Absent starts at the first line. |
+| `--offset <value>` | No | First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line. |
 | `--limit <value>` | No | How many lines to return from `offset`. Absent reads to the end. |
 
 </CommandTable>
@@ -1281,7 +1281,7 @@ sim files read <fileId> [options]
 | Option | Required | Description |
 | --- | --- | --- |
 | `--max-bytes <value>` | No | Optional ceiling on the source bytes fed to the parser, lowering but never raising the server limit. |
-| `--offset <value>` | No | First line to return, 1-based. Absent starts at the first line. |
+| `--offset <value>` | No | First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line. |
 | `--limit <value>` | No | How many lines to return from `offset`. Absent reads to the end. |
 
 </CommandTable>

apps/docs/openapi-v2-files-audit.json

@@ -790,11 +790,11 @@
             "name": "offset",
             "in": "query",
             "required": false,
-            "description": "First line to return, 1-based. Absent starts at the first line.",
+            "description": "First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line.",
             "schema": {
-              "description": "First line to return, 1-based. Absent starts at the first line.",
+              "description": "First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line.",
               "type": "integer",
-              "minimum": 1,
+              "minimum": 0,
               "maximum": 9007199254740991
             }
           },
@@ -1241,11 +1241,11 @@
             "name": "offset",
             "in": "query",
             "required": false,
-            "description": "First line to return, 1-based. Absent starts at the first line.",
+            "description": "First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line.",
             "schema": {
-              "description": "First line to return, 1-based. Absent starts at the first line.",
+              "description": "First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line.",
               "type": "integer",
-              "minimum": 1,
+              "minimum": 0,
               "maximum": 9007199254740991
             }
           },

apps/sim/app/api/v2/files/[fileId]/text/route.test.ts

@@ -112,6 +112,16 @@ describe('GET /api/v2/files/[fileId]/text', () => {
     expect(mocks.readText).not.toHaveBeenCalled()
   })
 
+  /**
+   * Callers used to zero-based paging send `offset=0` for the first window. It
+   * names no other line, so it reads from the first one instead of failing.
+   */
+  it('reads offset 0 as the first line', async () => {
+    const response = await GET(textRequest(`workspaceId=${WORKSPACE_ID}&offset=0&limit=2`), context)
+
+    expect(response.status).toBe(200)
+  })
+
   it('conceals a cross-tenant file as a missing file', async () => {
     mocks.readText.mockRejectedValueOnce(new NoWorkspaceAccessError())
 

apps/sim/lib/api/contracts/tools/file.ts

@@ -209,11 +209,12 @@ export const fileManageContentBodySchema = z
     fileId: fileIdSelectionSchema.optional(),
     fileInput: z.unknown().optional(),
     /**
-     * First line to return, 1-based. Applied to each selected file separately,
-     * because a selection can be several files and one running offset across
-     * them would depend on an ordering the caller cannot see.
+     * First line to return, 1-based; 0 also starts at the first line, so a
+     * zero-based caller's first window reads instead of failing. Applied to each
+     * selected file separately, because a selection can be several files and one
+     * running offset across them would depend on an ordering the caller cannot see.
      */
-    offset: z.number().int().min(1, 'offset starts at line 1').optional(),
+    offset: z.number().int().min(0, 'offset cannot be negative').optional(),
     /** How many lines to return from `offset`. Absent reads to the end. */
     limit: z.number().int().min(1, 'limit must be at least 1').optional(),
   })

apps/sim/lib/api/contracts/v2/files.ts

@@ -853,9 +853,11 @@ export const v2ReadFileTextQuerySchema = z
     offset: z.coerce
       .number()
       .int()
-      .min(1, 'offset starts at line 1')
+      .min(0, 'offset cannot be negative')
       .optional()
-      .describe('First line to return, 1-based. Absent starts at the first line.'),
+      .describe(
+        'First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line.'
+      ),
     limit: z.coerce
       .number()
       .int()

apps/sim/lib/internal/file/operations.test.ts

@@ -808,6 +808,21 @@ describe('file manage operations', () => {
       },
     })
   })
+  it('reads offset 0 as the first line of each selected file', async () => {
+    const response = await POST(
+      createMockRequest('POST', {
+        operation: 'content',
+        workspaceId: 'workspace-1',
+        fileId: 'file-1',
+        offset: 0,
+        limit: 1,
+      })
+    )
+    expect(response.status).toBe(200)
+    expect(await response.json()).toMatchObject({
+      data: { contents: ['content:file-1.txt'], lineRanges: [{ offset: 1, lineCount: 1 }] },
+    })
+  })
   it('does not bypass complete extraction limits with a raw-text fallback', async () => {
     vi.mocked(isSupportedFileType).mockReturnValueOnce(true)
     vi.mocked(parseBuffer).mockRejectedValueOnce(

apps/sim/lib/mothership/agent-cli/engines/file-read.test.ts

@@ -205,11 +205,28 @@ describe('content-aware files read augmentation', () => {
     }
   )
 
+  it('forwards --offset 0 to the text API instead of refusing it', async () => {
+    mockResolveWorkspaceFileReference.mockResolvedValue({
+      ...file,
+      name: 'notes.txt',
+      type: 'text/plain',
+    })
+    mocks.request.mockResolvedValue(textResponse('notes.txt'))
+    const result = await runEngine('files read', ['uploads/notes.txt'], runtime, {
+      offset: '0',
+      limit: '2',
+    })
+    expect(result.exitCode).toBe(0)
+    expect(mocks.request).toHaveBeenCalledWith(`/api/v2/files/${fileId}/text`, {
+      query: { workspaceId, offset: '0', limit: '2' },
+    })
+  })
+
   it('rejects incompatible flags and invalid byte/line bounds before lookup', async () => {
     for (const flags of [
       { render: true, limit: '2' } as const,
       { pages: '1', offset: '1' },
-      { offset: '0' },
+      { offset: '-1' },
       { offset: true } as const,
       { limit: true } as const,
       { 'max-bytes': true } as const,

apps/sim/lib/workspace-files/text-lines.test.ts

@@ -25,6 +25,12 @@ describe('file text line windows', () => {
         }
     }
   )
+  it('reads offset 0 as the first line and reports the 1-based offset it used', () => {
+    expect(sliceFileTextLines('one\ntwo\nthree', 0, 2, false)).toEqual({
+      text: 'one\ntwo',
+      lineRange: { offset: 1, lineCount: 2, totalLines: 3, totalLinesExact: true },
+    })
+  })
   it('reads a late window in a newline-dense document', () => {
     const result = sliceFileTextLines(
       `${'abc\n'.repeat(1_000_000)}tail needle\n`,

packages/sim-cli/src/generated/v2-api.ts

@@ -20276,7 +20276,8 @@ export const V2_OPERATIONS = {
       },
       offset: {
         kind: 'integer',
-        describe: 'First line to return, 1-based. Absent starts at the first line.',
+        describe:
+          'First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line.',
       },
       limit: {
         kind: 'integer',
@@ -20300,7 +20301,8 @@ export const V2_OPERATIONS = {
       },
       offset: {
         kind: 'integer',
-        describe: 'First line to return, 1-based. Absent starts at the first line.',
+        describe:
+          'First line to return, 1-based; 0 also starts at the first line. Absent starts at the first line.',
       },
       limit: {
         kind: 'integer',