بنقرة واحدة
spike
Throwaway experiments to validate an idea before building. Disposable by design.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
القائمة
Throwaway experiments to validate an idea before building. Disposable by design.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Karpathy-inspired guidelines: Think before coding, Simplicity first, Surgical changes, Goal-driven execution
Use when you have a written implementation plan to execute in a separate session with review checkpoints
Write an actionable markdown plan before implementation. Bite-sized tasks, exact paths, complete code.
Write clear Conventional Commits. Imperative subject, explain why not what, one logical change per commit.
Optimize based on measurement, never guesswork. Profile first, fix the real bottleneck, verify the win.
Refactor code without changing behavior. Small steps, green tests after every change, no mixing with feature work.
| name | spike |
| description | Throwaway experiments to validate an idea before building. Disposable by design. |
| conditional | true |
| paths | *.go,*.py,*.js,*.ts,*.rs,*.java |
Throwaway experiments to validate feasibility before committing to a real build. Spikes are disposable — throw them away once they've answered the question.
plan skill insteadBreak the idea into 2-5 independent feasibility questions:
| # | Spike | Validates | Risk |
|---|---|---|---|
| 001 | websocket-streaming | Given WS connection, when LLM streams, then chunks < 100ms | High |
| 002a | pdf-parse-libA | Given PDF, when parsed with libA, then text extractable | Medium |
| 002b | pdf-parse-libB | Same question, different library | Medium |
Order by risk. The spike most likely to kill the idea runs first.
write to create _spike_<name>.go or similarbash to run itAfter each spike, state clearly:
Delete spike files when done:
rm _spike_*.go