grantscope

Surface-quality loop — judge a user-facing screen the way Ben does (Clarity / Value-shown / Meaning / Aesthetic / Friction), collect ranked findings without fixing mid-audit, then fix the top ones with Ben in the loop and ship them. Seeded from the buyer-flow UX audit. Self-paced loop; exit is Ben's taste-check, not "when done". Use on /polish, "polish this", "audit the UX", "does this surface read as value", or before declaring a buyer-facing screen good.

Data-to-Goals Leverage Map — inventory the whole data estate, align it to the buyer-wedge + mission goals, and surface the highest-value LATENT connections we already have the data for but aren't exploiting. Connect/deepen, never widen (data widening is paused). Self-paced loop that appends to a living map. Use on /leverage, "what should we build", "find valuable connections", "what are we sitting on", "align our data to goals".

Buyer-wedge strategy guardrail — evaluate any proposed feature, data work, or product decision against the SE-registry strategy (free open registry for everyone; paid evidence + tender tools for buyers). Use when deciding whether to build something, when scoping a feature, when prioritising between work items, or when the user asks "should we build X" / "/wedge".

Lighthouse-buyer workflow — find and prepare the first paying government buyer for the SE evidence + tender-pack product. Mines se_buyer_prospects (buyers who already contract with registry SEs), shortlists VIC/SA buyers with social-procurement obligations, and assembles an evidence-led outreach pack. Use when the user says "/lighthouse", "find buyers", "prospect buyers", or wants outreach material for a procurement team.

Brainstorm and evolve the /graph visualization — storytelling, data layers, interaction design, and new graph modes