# Expert Roundtable

> Use before decomposition for non-trivial product, workflow, or agentic feature requests. Select a small relevant panel of experts, surface the main tensions, and produce a concise synthesis memo that shapes the ExecPlan.

- Skill: `moorage/expert-roundtable` (Agent Skill)
- Install (CLI): `npx skillmds@latest add moorage/expert-roundtable`
- Raw SKILL.md: https://api.skillmd.com/api/skills/moorage/expert-roundtable/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: moorage (https://skillmd.com/u/moorage)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/moorage/expert-roundtable

---


Use this skill when a request benefits from multiple expert lenses.

Good fits:
- user-facing feature work
- messaging, permissions, or trust-sensitive workflows
- agentic features involving model behavior, prompts, or evals
- architecture changes with operational consequences

Do not use it for:
- trivial bug fixes
- narrow mechanical edits

## Goal

Improve the framing of the request before planning begins.

The output is a concise synthesis, not a long roleplay transcript.

## Typical expert panel for this repo

- Security researcher
- UX expert
- Staff engineer or architect
- Reliability or operations engineer
- Applied ML or evals expert when prompts, routing, or model behavior matter

## Output format

Produce a concise memo with these sections:

### Expert panel
- [expert] — [why selected]

### What problem are we actually solving?
[short reframing]

### Roundtable highlights
- [expert]: [main concern + best recommendation]

### Key tensions
- [tension]

### Synthesis for decomposition
- goals that should drive the plan
- constraints that must shape the plan
- major risks or unknowns
- sequencing implications that must appear in the ExecPlan