05-spike — Claude Code skill: what it does, how to install
05-spike turns one blocking unknown into a bounded spike: frames the question, gathers evidence, writes the outcome.
MIT ★ 571 updated 2026-10-29
/plugin marketplace add ai-driven-dev/framework
/plugin install aidd-pm@aidd-frameworkWhat 05-spike does
The skill runs a spike on the single question that stands in the way of an estimate, a feasibility check or a design call. It takes either the question itself or a spike that already exists.
Work moves through actions. create qualifies the question and persists it, investigate collects evidence, conclude records the outcome and syncs the parent items. Only the file for the next action is opened, not the whole set. Evidence, decisions and assumptions sit apart from each other, and links back to sources are kept.
Nothing gets written until you approve it, or until you hand over authority up front. The question carries a stop condition and the skill stays inside it. General research and implementation fall outside its scope, so dropping it into an ordinary refactor buys you nothing.
What it helps with
- frame the question that blocks an estimate
- collect evidence for a contested assumption
- record the spike outcome and sync parent items
- keep evidence, decisions and assumptions apart
- resume a saved spike where it left off
Example prompts for the agent
- Frame a spike around whether the current schema can hold the new import format.
- Pick up the saved spike and gather evidence for the open question.
- Conclude the spike: write the outcome and update the parent items.
How to install 05-spike
1Plugin marketplace
/plugin marketplace add ai-driven-dev/framework/plugin install aidd-pm@aidd-frameworkThis installs the whole aidd-pm plugin with all its skills
2Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 05-spike -a claude-code3Manually: put the files in place
~/.claude/skills/05-spike/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/05-spike/
1Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 05-spike -a codex2Manually: put the files in place
.agents/skills/05-spike/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 05-spike -a cursor2Manually: put the files in place
.cursor/skills/05-spike/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install ai-driven-dev/framework plugins/aidd-pm/skills/05-spike2Manually: put the files in place
.github/skills/05-spike/SKILL.md1Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 05-spike -a gemini-cli2Platform CLI
gemini skills install https://github.com/ai-driven-dev/framework --path plugins/aidd-pm/skills/05-spikeManually: put the files in place
.devin/skills/05-spike/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 05-spike -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about 05-spike
How do I install 05-spike in Claude Code?
Run /plugin marketplace add ai-driven-dev/framework first, then /plugin install aidd-pm@aidd-framework.
What does 05-spike do?
It runs a spike on the question blocking an estimate, a feasibility check or a design call: qualify the question, collect evidence, record the outcome.
Is 05-spike meant for general research?
No. The question carries a stop condition and the skill does not go past it; general research and implementation are not what it is for.
Source text
SKILL.md · MIT
--- name: 05-spike description: Produces an evidence-bounded spike for an uncertainty blocking estimation, feasibility, or design. Use when the user wants to frame, investigate, resume, or conclude one. Not for general research or implementation. argument-hint: question | spike --- # Spike ```mermaid flowchart LR source([question or Spike]) --> create source -->|"already persisted"| investigate create -->|"save for later"| done create -->|"investigate now"| investigate --> conclude conclude -->|"inconclusive"| investigate conclude --> done([Spike]) ``` ## Actions Run the flow above. Read only the next action file. | Action | Does | | ----------- | ------------------------------------- | | create | qualify and persist | | investigate | collect evidence | | conclude | write outcome and sync parents | ## Transversal rules - Keep product and lifecycle decisions with the user. - Separate evidence, decisions, and assumptions. - Preserve source links and existing edits. - Ask natural questions; never expose actions, references, or unchanged state. - Require explicit approval or caller-provided bounded authority before any write. - Bound the question; never answer beyond its stop condition.
© AI-Driven Dev · MIT · ai-driven-dev/framework