investigation

Coordinate parallel subagent and Codex investigations to gather codebase evidence before planning.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/tasai-lab/fractal-dev-workflow --skill investigation-tasai-lab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: investigation
Source: https://github.com/tasai-lab/fractal-dev-workflow/tree/main/skills/investigation
Command: npx skills add https://github.com/tasai-lab/fractal-dev-workflow --skill investigation-tasai-lab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Before planning, teams often proceed without concrete evidence from the codebase or external sources. This skill ensures structured, evidence-based investigation to ground decisions.

Core Features & Use Cases

  • Parallel investigations by subagents and Codex to gather codebase evidence and external information.
  • Mandatory inventory and analysis steps to summarize existing implementations and identify gaps.
  • Supports a structured workflow from investigation planning to results integration, reducing blind spots before planning.

Quick Start

Coordinate parallel subagent investigations and Codex queries before drafting a plan.

Frequently Asked Questions about investigation

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

FAQPage Schema
How do I gather codebase evidence before planning a new feature?

To investigate a codebase for requirements, you deploy parallel subagents and Codex queries to extract real data. This enforces mandatory inventory and diff analysis steps to identify existing implementations and gaps before planning begins.

What is parallel evidence gathering and how does it work for software planning?

Parallel evidence gathering dispatches multiple subagents and Codex queries simultaneously to collect codebase and external data. It enforces mandatory inventory and diff analysis to verify findings before integrating results into a structured plan.

How do I validate external information and codebase gaps before writing a project plan?

You validate external information and codebase gaps by executing parallel subagent investigations and Codex queries. This enforces a mandatory inventory step to summarize implementations and performs diff analysis to verify findings before planning.

Can I use subagents and Codex to investigate a codebase in parallel?

Yes, you can coordinate subagents and Codex to investigate a codebase in parallel. This skill structures their simultaneous execution to gather evidence, enforces inventory and diff analysis to verify findings, and integrates results for planning.

What is the best way to ground software planning in real data instead of assumptions?

The best way to ground software planning in real data is to coordinate parallel evidence gathering using subagents and Codex. This enforces mandatory inventory and diff analysis to summarize implementations and verify findings before drafting a plan.

Why does software planning fail without structured codebase investigation?

Software planning fails without structured codebase investigation because teams proceed without concrete evidence, creating blind spots. This skill solves the problem by mandating parallel evidence gathering and diff analysis to verify findings before planning.