explore

Blend in-thread reasoning with multi-tier research to converge nascent ideas.

Updated May 3, 2026
One-click install
npx skills add https://github.com/fkberthold/loom --skill explore-fkberthold
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: explore
Source: https://github.com/fkberthold/loom/tree/main/skills/explore
Command: npx skills add https://github.com/fkberthold/loom --skill explore-fkberthold

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides a lightweight upstream exploration primitive to converge nascent ideas into a shared understanding by blending in-thread reasoning with multi-tier research across self, repo/docs, web, and peer-reviewed literature. It sits upstream of the design cycle, does not act as a bead or epic, and captures the outcome in a decision drawer for later promotion or rest.

Core Features & Use Cases

  • Lightweight, multi-tier exploration that blends Tier 1 (self) and Tier 2 (repo/docs) reasoning with optional Tier 3/4 deep-research when needed.
  • HYBRID loop: in-thread reasoning plus dispatched heavy rounds to surface robust grounding from web and scholarly sources.
  • Two user exits: REST to park a finding, PROMOTE to wire into /design-a-cycle, with grounding in the exploration drawer.
  • Seed drawers using templates and tag them as exploration to surface in searches.

Quick Start

Provide an nascent idea and trigger the exploration by issuing /explore <your-idea>.

Frequently Asked Questions about explore

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

FAQPage Schema
How do I explore a nascent idea before committing to a design cycle?▼

Multi-tier exploration converges nascent ideas into shared understanding by blending in-thread reasoning with research across self-reflection, repo context, web data, and peer-reviewed literature. It grounds concepts before committing to a design cycle.

What is the best way to ground early-stage concepts using repository context and web research?▼

A hybrid loop blends light in-thread reasoning with heavy-dispatch research rounds, pulling robust grounding from repository docs, web data, and peer-reviewed literature to validate early-stage concepts.

Can I park exploration findings instead of immediately promoting them to a design cycle?▼

Yes, you can use the REST exit to park findings in an exploration drawer for later rest, or use the PROMOTE exit to wire grounded findings directly into a design cycle.

Does this exploration process require a soundness gate before capturing findings?▼

No, the exploration process requires no soundness gate. It captures firm findings into a decision drawer using a hybrid loop of light in-thread work and heavy-dispatch research rounds.

How do I make exploration drawers discoverable in searches using templates?▼

Seed drawers using templates and tag them as exploration to surface them in searches. This organizes captured findings for potential promotion to design cycles or later rest.