loom-spike

loom-spike

Spikes and sketches are research-shaped workflows.

They are useful when the project needs bounded discovery before commitment.

What This Workflow Coordinates

• spike experiment framing
• sketch variant framing
• throwaway write-scope discipline
• evidence and null-result capture
• downstream routing after discovery

Use This Skill When

• a technical question needs a bounded experiment
• a design question needs a few concrete variants
• a throwaway prototype should inform a spec, plan, ticket, or wiki page
• rejected paths and null results should remain citable

Do Not Use This Skill When

• the intended behavior is already clear enough for a normal ticket
• the work needs production-quality implementation
• the artifact should become accepted explanation without critique or evidence

Spike Flow

question -> experiment matrix -> bounded throwaway child write scope -> evidence -> conclusions/null results -> downstream route

If the spike only reads, compares, sketches, or records observations, research and evidence may be enough. If the spike writes throwaway code or mutates the repository, create or tighten a ticket and use a Ralph packet with explicit cleanup expectations.

Record:

• question
• method
• experiment matrix
• child write scope and cleanup expectation
• evidence gathered
• conclusions
• null results or rejected paths
• recommended downstream owner

Sketch Flow

design question -> 2-3 variants -> screenshots or artifacts -> critique -> accepted wiki/spec updates

Sketches are the Loom adaptation of visual brainstorming. Use them when seeing a mockup, diagram, state flow, layout comparison, or spatial relationship would make a decision more honest than text alone.

Record:

• design question
• variants
• screenshots, prototypes, or other artifacts
• critique findings
• accepted behavior or explanation
• downstream spec, wiki, or ticket recommendation

If a harness or local tool helps produce visual artifacts, treat that tool as transport. It does not become a Loom layer. Preserve durable outputs in evidence and route accepted behavior or explanation to spec or wiki.

Done Means

• research owns conclusions and null results
• evidence owns observed artifacts
• any accepted behavior is routed to spec
• any accepted explanation is routed to wiki
• throwaway code is removed or explicitly contained

Read In This Order

Read immediately for spike or sketch work:

1. skills/loom-research/SKILL.md when recording experiment method, conclusions, rejected options, or null results.
2. skills/loom-evidence/SKILL.md when preserving screenshots, logs, artifacts, or observed outputs.

Then read conditionally:

3. skills/loom-tickets/SKILL.md and skills/loom-ralph/SKILL.md when the spike or sketch writes throwaway code or mutates repository files.
4. skills/loom-critique/SKILL.md when variants or experiment conclusions need adversarial review.
5. skills/loom-specs/SKILL.md when accepted behavior should become a contract.
6. skills/loom-wiki/SKILL.md when accepted explanation should become reusable.