sp-deep-research

Coordinates feasibility research, evidence packets, and disposable spikes before implementation planning.

Updated Jul 6, 2026
One-click install
npx skills add https://github.com/chenziyang110/launchdeck --skill sp-deep-research-chenziyang110
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sp-deep-research
Source: https://github.com/chenziyang110/launchdeck/tree/main/.claude/skills/sp-deep-research
Command: npx skills add https://github.com/chenziyang110/launchdeck --skill sp-deep-research-chenziyang110

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When a spec is written but key capabilities still lack a proven implementation path, teams risk planning on unverified assumptions. This Skill closes that gap by coordinating structured feasibility research, evidence collection, and disposable prototype spikes before implementation planning begins. ## Core Features & Use Cases - Multi-Track Research Orchestration: Decomposes feasibility questions into independent research tracks dispatched to subagents, each returning structured evidence packets with confidence levels and planning implications. - Disposable Demo Spikes: Builds isolated proof-of-concept spikes under research-spikes/ to validate uncertain implementation chains without touching production code. - Traceable Planning Handoff: Produces deep-research.md with stable IDs (CAP, TRK, EVD, SPK, PH) and an evidence quality rubric so /sp-plan can cite exactly which finding supports each design decision. - Use Case: Before planning a feature that depends on an unfamiliar third-party API, run this Skill to verify the API's constraints, spike a minimal integration, and hand the planner a proven approach with rejected alternatives documented. ## Quick Start Ask the agent to run /sp-deep-research with a description of the feasibility question or research tracks needed before planning handoff.

Frequently Asked Questions about sp-deep-research

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

FAQPage Schema
How do I validate feature feasibility before implementation planning?▼

Run /sp-deep-research after your spec is written but before /sp-plan. It decomposes feasibility questions into research tracks, gathers evidence from the repository and external sources, and produces a planning handoff with proven approaches and rejected options.

What is a research spike in spec-driven development?▼

A research spike is a small disposable prototype built to prove an uncertain implementation chain. This Skill isolates spikes under FEATURE_DIR/research-spikes/ and records hypothesis, setup, expected and actual results, plus what the spike does not prove.

Does sp-deep-research modify production source code?▼

No. The Skill is research-only and explicitly forbids editing production source, tests, migrations, or release config. Subagents may only return evidence packets or write to their assigned disposable spike directory.

When should I skip the deep research phase?▼

Skip it for minor adjustments to already-proven capabilities or when alignment.md marks capabilities as Not needed or Completed. The Skill still writes a lightweight deep-research.md with Status: Not needed to preserve the workflow record.

Why was my subagent evidence packet rejected?▼

Packets are rejected when paths_read is empty, the finding is speculative or unanswered, or the subagent edited production files. Rejected tracks get one retry with clarified instructions before being marked blocked and escalated.

What project setup does sp-deep-research require?▼

It requires a spec-kit project structure with a .specify/ directory, an active feature directory containing spec.md and alignment.md, and workflow-state.md for phase tracking. The check-prerequisites script resolves FEATURE_DIR before research begins.