# Lets Talk About It

> One-pass interrogation and coding guardrails. Use to stress-test a plan, design, code change, or domain term against local code/docs while enforcing explicit assumptions, simplicity, surgical edits, and verifiable success criteria. Also guides a longer feature discussion and turns the settled feature into a copy-ready prompt for a separate implementation session with a chosen sub-agent model roster. Use to explore, narrow, summarize, or hand off a substantial feature without implementing it.

- Skill: `christopheralphonse/lets-talk-about-it` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx get-skill add christopheralphonse/lets-talk-about-it`
- Raw SKILL.md: https://api.skillmd.com/api/skills/christopheralphonse/lets-talk-about-it/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: ChristopherAlphonse (https://skillmd.com/u/christopheralphonse)
- Updated: 2026-11-16
- Page: https://skillmd.com/skills/christopheralphonse/lets-talk-about-it

---


# Lets Talk About It

Stress-test before committing. Optimize for one useful pass, not exhaustive questioning.

## Operating Contract

- Inspect local evidence first: `CONTEXT.md`, `CONTEXT-MAP.md`, ADRs, specs, plans, task lists, configs, tests, and nearby source.
- Ask at most one question at a time, only when the answer changes scope, architecture, risk, terminology, or verification.
- State assumptions, contradictions, tradeoffs, and unresolved decisions explicitly.
- Prefer the simplest viable option. Do not add speculative features, abstractions, configurability, or future-proofing.
- Keep edits surgical: touch only requested files/behavior, match local style, and clean up only unused code introduced by your own changes.
- Define success criteria before implementation, then verify with the narrowest meaningful check.

## One-Pass Flow

1. **Frame:** Restate the goal, known context, assumptions, and non-goals.
2. **Check evidence:** Compare the plan or claim against local docs and code. Cite files for important findings.
3. **Interrogate:** Walk only the relevant branches: problem, users, requirements, terminology, architecture, data/API flow, security, performance, reliability, tests, deployment.
4. **Decide:** Recommend the simplest defensible path and name rejected alternatives.
5. **Verify:** List concrete checks: tests, commands, browser steps, metrics, or review criteria.

## Output

Return only what the task needs:

- Recommendation
- Assumptions / open questions
- Scope and non-goals
- Risks / tradeoffs
- Verification checklist
- `CONTEXT.md` or ADR updates, only when a durable domain term or hard-to-reverse decision was actually resolved. Follow [CONTEXT-FORMAT.md](CONTEXT-FORMAT.md) for `CONTEXT.md`/`CONTEXT-MAP.md` and [ADR-FORMAT.md](ADR-FORMAT.md) for ADRs.

## Feature Handoff

When the user wants to explore a substantial feature over several turns and hand it to a separate implementation session, follow `FEATURE-ORCHESTRATOR.md`. It does not implement in this session.

