plan-grill-and-save

Interviews users to resolve requirements and writes self-contained implementation plans saved via write_saved_plan_file.

2|Updated Apr 2, 2026
One-click install
npx skills add https://github.com/nseng-ai/ns --skill plan-grill-and-save-nseng-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plan-grill-and-save
Source: https://github.com/nseng-ai/ns/tree/main/skills/incubating/branch-context/plan-grill-and-save
Command: npx skills add https://github.com/nseng-ai/ns --skill plan-grill-and-save-nseng-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Implementation plans often fail because requirements are ambiguous or the plan depends on conversational context that a fresh coding-agent session cannot see. This Skill resolves material requirements through targeted questions, then produces a self-contained, reviewed plan saved through the repository's existing Saved Plan writer. ## Core Features & Use Cases - Requirements Interview: Loads the installed grilling skill to ask only judgment-dependent questions that materially change the plan, skipping facts already discoverable in the repository. - Self-Contained Plan Writing: Produces a Markdown plan with goals, evidence, decisions, ordered implementation steps, validation guidance, and risks that a fresh downstream agent can execute without prior conversation. - Deterministic Saving: Calls write_saved_plan_file exactly once, letting the existing writer own slug derivation, collision refusal, and storage. - Use Case: Before starting a multi-file refactor, run this workflow to grill the author on design tradeoffs, then hand the resulting Saved Plan to a new coding-agent session for implementation. ## Quick Start Ask the agent to run the plan-grill-and-save workflow to interview you about requirements and save a reviewed implementation plan for your intended change.

Frequently Asked Questions about plan-grill-and-save

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create an implementation plan for a coding agent?

Run this workflow with a planning target inferred from your request, conversation, or repository. It interviews you on material decisions, writes a self-contained Markdown plan, and saves it via write_saved_plan_file for a fresh agent session to execute.

What questions does the planning interview ask?

It asks only questions requiring user judgment that materially change the plan, following the installed grilling skill. It never asks about facts available in the repository or routine test coverage questions unless validation is itself a product requirement.

Does plan-grill-and-save work without the grilling skill installed?

No. The grilling skill is a required dependency that owns the interview procedure. If the harness cannot load it, the workflow stops and reports the missing dependency rather than approximating the interview.

Can I save the plan with a custom filename or shell command?

No. The plan must be saved by calling write_saved_plan_file exactly once with the full Markdown content. The existing writer owns slug derivation, collision refusal, and storage; shell-based writes are explicitly forbidden.

What happens after the plan is saved?

The workflow reports the saved-plan evidence including file path, repository, source branch, and slug, then stops. It does not create branches, start implementation, or invoke other branch-context workflows.