XLSX Skill

Thin wrapper that routes XLSX creation, editing, and analysis requests to the sheet-agent, which manipulates spreadsheets using openpyxl and pandas via Bash.

Context Pointers

Reference (do not load eagerly):

• Path: .claude/context/formats/subagent-return.md
• Purpose: Return validation
• Load at: Subagent execution only

Note: This skill is a thin wrapper. Context is loaded by the delegated agent, not this skill.

Trigger Conditions

This skill activates when:

Direct Invocation

• User explicitly runs /sheet command
• User runs /edit command with a .xlsx file

Implicit Invocation (during task implementation)

When an implementing agent encounters any of these patterns:

Plan step language patterns:

• "Create spreadsheet [file].xlsx with [content]"
• "Create xlsx [file] with [columns/data]"
• "Edit [file].xlsx to [change]"
• "Edit spreadsheet [file] to [change]"
• "Add formulas to [file].xlsx"
• "Add formatting to [file].xlsx"
• "Update [file].xlsx with [data]"
• "Analyze [file].xlsx for [purpose]"

File extension detection:

• Target file has extension: .xlsx, .xlsm
• Operation type: create, edit, update, modify, analyze, add formulas, add formatting

Task description keywords:

• "XLSX creation"
• "spreadsheet creation"
• "xlsx editing"
• "spreadsheet editing"
• "add formulas"
• "add formatting"
• "xlsx analysis"

When NOT to Trigger

Do not invoke for:

• Spreadsheet-to-table conversion (use skill-filetypes-spreadsheet for /table)
• Simple CSV/TSV reading (use Read tool directly)
• Grant budget creation (use skill-budget for /budget)
• PDF/DOCX operations (use other filetypes skills)
• Presentation operations (use skill-presentation)

Execution

1. Input Validation

Validate required inputs:

• file_path - Must be provided; for edit/analyze must exist; for create mode may not exist yet
• instruction - Natural language instruction describing the desired operation (required)
• mode - One of: "create", "edit", "analyze" (inferred from context and file existence)

# For edit mode: validate source exists and is xlsx/xlsm
if [ "$mode" = "edit" ]; then
  if [ ! -f "$file_path" ]; then
    return error "XLSX file not found: $file_path"
  fi
  if [[ "${file_path##*.}" != "xlsx" && "${file_path##*.}" != "xlsm" ]]; then
    return error "Source must be a .xlsx or .xlsm file: $file_path"
  fi
fi

# For analyze mode: validate source exists
if [ "$mode" = "analyze" ]; then
  if [ ! -f "$file_path" ]; then
    return error "File not found for analysis: $file_path"
  fi
fi

# For create mode: validate xlsx extension
if [ "$mode" = "create" ]; then
  if [[ "${file_path##*.}" != "xlsx" && "${file_path##*.}" != "xlsm" ]]; then
    return error "New file must have .xlsx or .xlsm extension: $file_path"
  fi
fi

2. Context Preparation

Prepare delegation context:

{
  "file_path": "/absolute/path/to/spreadsheet.xlsx",
  "instruction": "Create a budget tracking spreadsheet with categories, monthly columns, and SUM formulas",
  "mode": "create",
  "metadata": {
    "session_id": "sess_{timestamp}_{random}",
    "delegation_depth": 1,
    "delegation_path": ["orchestrator", "xlsx", "skill-sheet"]
  }
}

3. Invoke Agent

CRITICAL: You MUST use the Agent tool to spawn the sheet-agent.

Required Tool Invocation:

Tool: Agent (NOT Skill, NOT Plan)
Parameters:
  - subagent_type: "sheet-agent"
  - prompt: [Include file_path, instruction, mode, metadata]
  - description: "{mode} {file_path}: {instruction}"

DO NOT use Skill(sheet-agent) - this will FAIL. Agents live in .claude/agents/ or extension agent directories, not .claude/skills/. The Skill tool can only invoke skills from .claude/skills/.

The agent will:

• Detect available tools (openpyxl, pandas)
• Execute the requested operation (create, edit, or analyze)
• Validate the output file
• Return standardized JSON result

4. Return Validation

Validate return matches subagent-return.md schema:

• Status is one of: created, edited, analyzed, partial, failed
• Summary is non-empty and <100 tokens
• Artifacts array present with output file path
• Metadata contains session_id, agent_type, delegation info

Status value rules:

• created - New file successfully created
• edited - Existing file successfully modified
• analyzed - File read and analysis returned
• partial - Some operations succeeded, others failed
• failed - No operations applied, error occurred
• NEVER use "completed" (triggers Claude stop behavior)

5. Return Propagation

Return validated result to caller without modification.

Return Format

See .claude/context/formats/subagent-return.md for full specification.

Expected successful return:

{
  "status": "created",
  "summary": "Created budget.xlsx with 4 category rows, monthly columns Jan-Dec, SUM formulas for totals, and currency formatting.",
  "artifacts": [
    {
      "type": "implementation",
      "path": "/absolute/path/to/budget.xlsx",
      "summary": "XLSX workbook with formulas and formatting"
    }
  ],
  "metadata": {
    "session_id": "sess_...",
    "agent_type": "sheet-agent",
    "delegation_depth": 2,
    "delegation_path": ["orchestrator", "xlsx", "skill-sheet", "sheet-agent"],
    "tool_used": "openpyxl",
    "mode": "create",
    "sheets": 1,
    "rows": 6,
    "columns": 14
  },
  "next_steps": "Open spreadsheet to verify formulas and formatting"
}

Error Handling

Input Validation Errors

Return immediately with failed status if file not found, not a .xlsx/.xlsm, or instruction is empty.

Unsupported Format

Return failed status if file extension is not .xlsx or .xlsm (for create/edit). Suggest using /convert or /table for other spreadsheet operations.

Agent Errors

Pass through the agent's error return verbatim.

Tool Not Available

Return failed status with installation instructions for openpyxl and pandas.

MUST NOT

• Run the postflight step (git commit, status update) -- that is the command's responsibility
• Modify the return from sheet-agent before propagating
• Load context files eagerly -- only reference them when needed
• Use the Skill tool to invoke sheet-agent (use Agent tool instead)