08-three-amigos — Claude Code skill: what it does, how to install
08-three-amigos evaluates epics and user stories through product, delivery, or quality perspectives, then reconciles three separate reports before backlog edits.
MIT ★ 571 updated 2026-10-29
/plugin marketplace add ai-driven-dev/framework
/plugin install aidd-pm@aidd-frameworkWhat 08-three-amigos does
The skill brings the Three Amigos review practice into your agent workflow. It handles two distinct actions: assess and reconcile. Running assess applies a single analytical viewpoint to an epic or story — examining product value, delivery feasibility, or testability. Running reconcile takes three separate role reports provided by the user and compiles their findings, highlighting conflicting requirements and unaddressed questions.
Only one action runs per prompt invocation. The skill never writes changes to disk or creates tasks automatically: every output is strictly a proposal grounded in cited text.
What it helps with
- evaluate user stories through a delivery or QA perspective
- spot requirement conflicts between product and engineering inputs
- merge three role reports into a unified findings summary
- generate clarifying questions before updating backlog items
Example prompts for the agent
- Assess
story-checkout.mdthrough the quality assurance lens - Reconcile these three role reports and flag any conflicts:
reports/ - Run assess on
epic-billing.mdfocusing strictly on delivery
How to install 08-three-amigos
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 08-three-amigos -a claude-code3Manually: put the files in place
~/.claude/skills/08-three-amigos/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/08-three-amigos/
1Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 08-three-amigos -a codex2Manually: put the files in place
.agents/skills/08-three-amigos/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 08-three-amigos -a cursor2Manually: put the files in place
.cursor/skills/08-three-amigos/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install ai-driven-dev/framework plugins/aidd-pm/skills/08-three-amigos2Manually: put the files in place
.github/skills/08-three-amigos/SKILL.md1Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 08-three-amigos -a gemini-cli2Platform CLI
gemini skills install https://github.com/ai-driven-dev/framework --path plugins/aidd-pm/skills/08-three-amigosManually: put the files in place
.devin/skills/08-three-amigos/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add ai-driven-dev/framework --skill 08-three-amigos -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 08-three-amigos
How to install 08-three-amigos in Claude Code?
First, register the marketplace using /plugin marketplace add ai-driven-dev/framework. Then install the package by running /plugin install aidd-pm@aidd-framework.
What is the purpose of 08-three-amigos?
It reviews stories or epics using a single analytical lens, or combines three pre-written role reports to reveal disagreements and open questions.
Does 08-three-amigos edit backlog files directly?
No. The skill rules strictly forbid file mutation, persistence, and spawning: it only produces analytical proposals and cited questions for the caller.
Source text
SKILL.md · MIT
--- name: 08-three-amigos description: Assesses an Epic or Story through one product, delivery, or quality lens, then reconciles three caller-supplied reports. Use when the user wants to refine one before a backlog change. Not for spawning or writing. argument-hint: artifact | reports --- # Three Amigos ```mermaid flowchart LR target([Epic or Story]) --> assess --> report([one role report]) reports([three role reports]) --> reconcile --> result([findings, conflicts, questions]) ``` ## Actions Run the flow above. Read only the next action file. | Action | Does | | --------- | ------------------------------------------------ | | assess | apply one requested lens to an Epic or Story | | reconcile | reconcile three independent, caller-supplied reports | ## Transversal rules - Run one route per invocation. - Never spawn, delegate, persist, or mutate. - Treat roles as analytical lenses, not human authorities. - Ground every finding and question in cited evidence; never invent a decision. - Return proposals only; the caller decides what to do with them.
© AI-Driven Dev · MIT · ai-driven-dev/framework