researcher

Skill

Skill

Skill

v0.3.0 — Turn how you actually work into a structured process model. A guided interview extracts what people forget to write down — triggers, data movement, decision criteria, exceptions, approval authority — then rates every step Automate, Monitor, or Human so you know where automation is safe and where a human must stay in the loop. The ratings are a reviewed draft, not a safety certification, until a stakeholder signs off. Discover sweeps your recurring work into a process inventory when you can't yet name your processes; capture models a process you already run; design models a new one you don't yet, proposing an intended flow grounded in your real constraints and nearest analog, never invented. Standalone.

v0.2.1 — Set business goals that survive contact with reality, then actually operate against them — not a vision board you abandon by February. A six-stage Setup Arc takes you from what game you're playing to goals stress-tested before launch; a five-cadence Ongoing Arc (daily writing to annual vision check) keeps them alive. A hard cap of three active goals and a critic that red-teams every goal before you commit keep it honest. Standalone.

v0.3.0 — Start the day knowing what actually moved in your world, without reading the whole internet to find it. It scans news, industry, research, policy, and science, triages it down to the few items worth your attention (keyed to a relevance context you set once), and writes them up as a clean, self-contained page you can read anywhere or forward. Self-contained; runs on its own.

v1.2.0 — Get a professional-grade photo from a plain-language description, without knowing what a key light or an 85mm prime is. A guided workflow turns your scene into a physics-aware Nano Banana Pro prompt — camera, lens, lighting, and color grade chosen from a curated reference library — then generates the image via the Gemini API at the right aspect ratio for print, social, or hero use. /photo-setup handles the API key and an optional default house style; batch and variations modes cover shot lists and multiple takes. Standalone.

v0.2.0 — End the week knowing where every decision landed and what's owed, even when five meetings all touched the same thing. It turns a stream of transcripts into one living weekly round-up of action items, cross-meeting threads, and a forward watch list. Pulls from Read.ai, Fireflies, or Granola via MCP; works on any dropped transcript otherwise. Self-contained.

v0.4.1 — Think a hard problem all the way through and come out with a strategy you can defend, not a blank framework template. It walks one problem through a repeatable loop (Define, Frame, Analyse, Insight, Synthesise, Story, Move), puts the right framework from a 70-framework library in front of you at each step, and pressure-tests your reasoning before you commit. Produces a working record and a clean reader-facing brief. Standalone.

v0.1.0 — Make sense of a confusing situation: what's actually happening, whether your own thinking is off, and what to do about it. It names the bias, fallacy, persuasion or manipulation tactic, or strategic move at work from a 243-pattern library, carefully (it separates real gaslighting from an ordinary disagreement before handing you a heavy label). Five skills: identify, explain, practice, decide, spar. Talks like a counselor, not a textbook. Standalone.