motleyai/update-table-block

Update Table Block

Create or modify table blocks using the update_table_block MCP tool. Table blocks support variable substitution, LLM-generated tables, and size constraints via target_shape.

How It Works

1. Call update_table_block with a user_prompt template
2. The block resolves immediately — variables are substituted and (optionally) LLM generates the table
3. The resolved table content is returned inline in the tool response

Returns: The resolved table content as markdown.

target_shape Format

Controls the expected table dimensions as [rows, cols]. Each dimension can be:

| Value | Meaning | Example | |-------|---------|---------| | null | No constraint | [null, 3] — any rows, exactly 3 columns | | int | Exact count | [5, 3] — exactly 5 rows and 3 columns | | [min, max] | Range | [[3, 8], 4] — 3-8 rows, exactly 4 columns |

Examples:

• [5, 3] — exactly 5 rows, 3 columns
• [null, 4] — any number of rows, 4 columns
• [[1, 10], null] — 1-10 rows, any columns
• [[3, 8], [2, 5]] — 3-8 rows, 2-5 columns

Row count does NOT include the header row.

Three Patterns

Pattern A: Raw Query Output

Pass a table mode query result directly as the table. No LLM needed.

# First, create a table-mode query:
update_query_block(
    parent_location={doc_id: 42, slide_name: "Data", block_name: "data_table"},
    query_name="monthly_data",
    prompt="Monthly revenue and order count for last 12 months",
    mode="table",
    model_name="orders"
)

# Then set the table template to just reference the query:
update_table_block(
    location={doc_id: 42, slide_name: "Data", block_name: "data_table"},
    user_prompt="{monthly_data}",
    target_shape=[[1, 12], 3]
)

Pattern B: Markdown Template with Variables

Build a table template with inline variable substitution.

update_table_block(
    location={doc_id: 42, slide_name: "KPIs", block_name: "kpi_table"},
    user_prompt="| Metric | Value |\n|--------|-------|\n| Revenue | {currency(total_revenue)} |\n| Customers | {integer(active_customers)} |\n| Growth | {percent(revenue_growth)} |",
    target_shape=[3, 2]
)

Pattern C: LLM-Generated from Data

Let the LLM build the table from query data.

update_table_block(
    location={doc_id: 42, slide_name: "Analysis", block_name: "summary_table"},
    user_prompt="Create a summary table from this data:\n\n{detailed_data}\n\nShow the top 5 items by revenue with columns: Name, Revenue, Growth %, Status",
    call_llm=true,
    target_shape=[5, 4]
)

Template Syntax

Same as text blocks — supports {variable}, {Slide::Block}, arithmetic, and formatting functions.

See variable-reference-syntax.md for full details.

Workflow

Every reference in the template must exist before you call update_table_block. The block resolves immediately, so any {variable} that can't be found will cause a resolution failure.

• {query_name} → call update_query_block first
• {Slide::Block} → the referenced slide/block must already have content
• {context_var} (e.g. {end_month}, or any source-specific key from get_doc_variables) → no setup needed if get_doc_variables lists it

Create query blocks first, then set the table template:

1. Create data queries:

   update_query_block(
       parent_location={doc_id: 42, slide_name: "Performance", block_name: "perf_table"},
       query_name="region_data",
       prompt="Revenue and customer count by region, top 5 regions by revenue",
       mode="table",
       model_name="sales"
   )
   

2. Set the table template:

   update_table_block(
       location={doc_id: 42, slide_name: "Performance", block_name: "perf_table"},
       user_prompt="{region_data}",
       target_shape=[5, 3]
   )
   

Examples

Simple Data Table

update_table_block(
    location={doc_id: 42, slide_name: "Revenue", block_name: "revenue_table"},
    user_prompt="{quarterly_revenue}",
    target_shape=[4, 3]
)

KPI Scorecard

update_table_block(
    location={doc_id: 42, slide_name: "Scorecard", block_name: "scorecard_table"},
    user_prompt="| KPI | Target | Actual | Status |\n|-----|--------|--------|--------|\n| Revenue | {currency(target_revenue)} | {currency(actual_revenue)} | {revenue_status} |\n| Users | {integer(target_users)} | {integer(actual_users)} | {users_status} |\n| NPS | {target_nps} | {actual_nps} | {nps_status} |",
    target_shape=[3, 4]
)

LLM-Summarized Table

update_table_block(
    location={doc_id: 42, slide_name: "Executive View", block_name: "exec_table"},
    user_prompt="Based on the following monthly performance data, create a table summarizing each quarter with columns: Quarter, Revenue, Trend, Key Highlight.\n\n{monthly_data}\n\nKeep highlights to one short sentence each.",
    call_llm=true,
    target_shape=[4, 4]
)

Pivoted Cross-Tab

# Create a pivoted query first:
update_query_block(
    parent_location={doc_id: 42, slide_name: "Matrix", block_name: "matrix_table"},
    query_name="product_region_matrix",
    prompt="Revenue by product and region",
    mode="table",
    pivot_dimension="region",
    model_name="sales"
)

# Display it:
update_table_block(
    location={doc_id: 42, slide_name: "Matrix", block_name: "matrix_table"},
    user_prompt="{product_region_matrix}",
    target_shape=[null, null]
)

Overflow Handling

If the resolved table exceeds the element's size on the slide, the tool returns an overflow error. To fix:

• Reduce rows: Use target_shape with a lower row count, or add "top N" to your query prompt
• Reduce columns: Select fewer measures/dimensions in the query
• Shorten content: Use abbreviations or shorter labels
• Use LLM: Set call_llm=true and instruct it to be concise

Error Handling — behavior_if_query_fails

Same as text blocks:

• "drop_slide" — Slide is silently skipped during export
• "fail_resolution" — Resolution raises an error
• null (default) — Keeps existing behavior

Related Skills

• For creating query blocks: see the update-query-block skill
• For expression syntax: see variable-reference-syntax.md
• For context variables: see resolution-context.md