Interactive onboarding for a hill-climbing optimization project. Use this skill whenever the user wants to start a long, iterative effort to optimize a measurable score, minimize a loss, maximize accuracy, or beat a perf target. Trigger on phrases like "set…
Run the hill-climbing loop autonomously: spawn subagents that invoke the `hillclimb-execute` → `hillclimb-verify` → (when stuck) `hillclimb-brainstorm` skills, iteration after iteration, until a stop condition fires. Uses git for per-iteration checkpoints…
End-to-end hill-climbing workflow: onboard the project if it isn't set up yet, then run the autonomous loop. Use when the user wants to set up and iterate in one invocation. Triggers: "run the hillclimb skill", "set up and iterate on X", "start a…
Refresh the idea pool of a hill-climbing project. Reads the run history and current ideas, detects whether progress has stalled, then generates new candidate directions, biased toward exploration when stuck, toward exploitation when a recent run improved…
Execute one iteration of a hill-climbing project: pick the next idea from the pool, plan it, do the work, log a run. Use this skill whenever the user has an active project with `.hillclimb/state.html` and wants to make progress. Trigger on phrases like "run…
Verify the most recent execution run in a hill-climbing project. Reads `.hillclimb/state.html`, runs the project's verifier script, and updates the dashboard with the result. Use this skill whenever the user finishes an `hillclimb-execute` and wants to know…