Guides spike experiments to an explicit toss, keep, or promote outcome.
npx skills add https://github.com/scoots31/engineering-playbook --skill land-scoots31
Please help me install this Agent Skill. Skill: land Source: https://github.com/scoots31/engineering-playbook/tree/main/skills/land Command: npx skills add https://github.com/scoots31/engineering-playbook --skill land-scoots31
High-intent search queries and answers about installing and using this skill.
Run a landing conversation that asks directly whether to toss, keep, or promote the spike. Toss archives the code and records the learning, keep freezes it as a personal tool after one hygiene pass, and promote hands it to the full build framework with a brief.
Stop treating the spike code as a foundation and prepare a handoff brief covering what it proved, what exists versus what is missing, scope signals from graduation triggers, and known hacks that must not survive promotion.
Do one hygiene pass: rename temp files, add a one-line header explaining what it does and how to run it, remove hardcoded secrets, and re-run it end-to-end. Then freeze it and resist further cleanup.
Spikes silently become shipped products when no explicit landing decision happens. The fix is a mandatory end-of-spike conversation that forces a toss, keep, or promote choice, with no deferral allowed.
Toss when the spike did not work or worked but is not worth maintaining. The outcome is the learning: summarize what was tried, what happened, and what it means, record it as a spike memory, then archive or delete the code.
product-on-purposeDocument time-boxed spike results with findings and recommendations.
peterbamuhigireCollect client intake data and generate a 12-section strategy brief.
hagajiExtract reusable patterns from sessions into structured memory files.
MarbleSodasBuild disposable prototypes to validate uncertain ideas with evidence-backed verdicts.
