hill-climb
Run a bounded, evidence-driven improvement loop for code, prompts, skills, websites, or generated media. Invoke only when the user explicitly asks for /hill-climb or a measured improvement loop. Freeze a sensitive evaluation, measure baseline noise, change one hypothesis at a time, keep or revert from evidence, and confirm the winner on untouched cases.
- Version
- 0.3.0
- License
- MIT
- Compatibility
- Harness independent. Python 3.11+ for the optional offline comparison helper. Actual experiments require the project's runner and relevant runtime or media tools. Typed contract tools require Python 3.11+ and the managed command-schema runtime.
Pinned to revision 92f99ffdb0bc, so it is the text this page describes rather than whatever the author pushed since.
Files
- skills/hill-climb/SKILL.md
- skills/hill-climb/LICENSE
- skills/hill-climb/commands/help.md
- skills/hill-climb/commands/model.md
- skills/hill-climb/commands/native_compare.md
- skills/hill-climb/commands/verify.md
- skills/hill-climb/commands/version.md
- skills/hill-climb/references/distilled-sources.md
- skills/hill-climb/references/execution.md
- skills/hill-climb/references/instrument-and-loop.md
- skills/hill-climb/references/measurement.md
- skills/hill-climb/references/operating-guide.md
- skills/hill-climb/references/sources.md
- skills/hill-climb/references/storage.md
- skills/hill-climb/scripts/contract.json
- skills/hill-climb/scripts/hill_climb.py
- skills/hill-climb/scripts/json_contract.py
- skills/hill-climb/scripts/normalization.json
- skills/hill-climb/scripts/package_runtime.py
- skills/hill-climb/scripts/run.py
- skills/hill-climb/scripts/tkstack_state.py
- skills/hill-climb/scripts/workflow.cases.json
- skills/hill-climb/scripts/workflow.lock.json
- skills/hill-climb/templates/plan.json
Every link opens the file at its source, pinned to the revision this page describes.