Rendering And Dashboard Prompts

Example 1: Render a peptide table

User prompt

Render the peptide table for this candidate file as HTML so I can browse it.

What the MCP should do

  • call render_peptide_table

Expected output shape

  • HTML output path
  • no scoring unless a scored table is explicitly merged

Example 2: Build an ID dashboard

User prompt

Build an ID dashboard for the top 1000 predictions and include retention time metadata from this Feather file.

What the MCP should do

  • call build_id_dashboard
  • use top_n=1000
  • pass spectra_path

Expected output shape

  • summary JSON path
  • enriched CSV path
  • dashboard HTML path

Example 3: Render a batch of spectra

User prompt

Render observed, annotated, and mirror plots for these five selected PSMs.

What the MCP should do

  • prepare a small batch manifest
  • call render_spectra_batch

Expected output shape

  • output directory with comparison/, annotated/, and observed/
  • render status file

Example 4: Ask for a rendering mode recommendation

User prompt

For this peptide, should I use the mirror plot, annotated observed spectrum, or plain observed spectrum?

What the MCP should do

  • explain the differences between the three visualization modes
  • recommend one based on the user’s goal

Expected output shape

  • short rationale
  • no rendering unless requested