goodthinking

Context-blind agents

Skills spawn subagents without conversation context, ensuring fresh perspectives that aren't anchored to your framing.

Friction preservation

Dissent and disagreement are preserved, and conflicting views are presented as-is for you to resolve.

Sharpen a problem statement

Creates an initial framing of your intent, then spawns a context-blind agent to evaluate it for precision, clarity, completeness, and scope. If gaps are found, it asks targeted clarifying questions. A second agent checks whether the problem could be better framed at a different level of abstraction, surfacing alternative framings you may not have considered.

Break a problem into parts

Proposes an initial breakdown, then checks to see if the components are mutually exclusive, collectively exhaustive, at uniform abstraction levels, and actionable.

Stress test decisions under adversarial scrutiny

Spawns two parallel agents that argue for and against your decision or hypothesis. The agents collaborate to show you which parts of your thinking are solid and uncontested, and where there is tension.

Surfaces concerns you may have missed

Spawns a context-blind agent to challenge your decision by flagging any hidden assumptions, practical risks, opportunity costs, and second-order effects you may have missed. Important, relevant, and actionable concerns are presented back to you.

Check whether you've explored all options

A context-blind agent checks to see if your option set is complete. It evaluates options against seven integrity criteria: missing options, dimensional collapse, description asymmetry, criteria rigging, false exclusivity, scope mismatch, and constraint fabrication.

Generate diverse ideas using agents with different perspectives

Frames the problem as "How Might We" questions, then selects 2–4 different lenses designed to push thinking in different directions. Each lens gets its own parallel agent that generates ideas from that perspective. Results are consolidated and presented back to you.

Consolidate raw ideas into themes

Takes a raw list of ideas, deduplicates them, clusters similar items, and groups them into themes. It validates that the themes are MECE, actionable, and collectively complete.