spike-wrap-up

To convert spike research into reusable project skills, package the completed spike outputs into a deployable skill file at .agents/skills/<name>/SKILL.md. This captures the findings for automatic loading in future similar work sessions.

Preserving spike findings for future tasks involves wrapping the research documentation into a durable, project-local skill. The original spike files remain intact for forensics while the new skill references the originating spike and appends a decision log entry.

Yes, you can automate the wrap-up workflow for multiple completed spikes. The process supports a repeatable workflow that converts each spike's outputs into deployable skills, reducing manual handoffs and setup time across the project.

Creating a project-local skill requires the completed spike output files: SCOPE.md, research/*.md, and RECOMMENDATION.md. These files provide the source material packaged into the new skill file under .agents/skills/<name>/SKILL.md.

Packaging spike outputs into a skill instead of keeping standalone markdown files ensures the research auto-loads during future similar work. This turns temporary research into durable tooling, eliminating the need to manually locate and reference past findings.

You should not use a spike wrap-up workflow if the spike is incomplete or lacks the required outputs like RECOMMENDATION.md. The process applies specifically after a spike completes and you want to convert its findings into durable tooling.