One-click install
npx skills add https://github.com/marcmunoz-uno/jailbreak --skill deep-dive-marcmunoz-uno
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-dive
Source: https://github.com/marcmunoz-uno/jailbreak/tree/main/skills/deep-dive
Command: npx skills add https://github.com/marcmunoz-uno/jailbreak --skill deep-dive-marcmunoz-uno

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Deep-dive helps when you need both a trustworthy causal explanation of what happened and a precise set of requirements for what to do next, without losing context between separate investigation and requirements steps.

Core Features & Use Cases

  • 3-lane causal trace: Runs three parallel investigations (implementation/code path, config/environment/orchestration, and measurement/artifact mismatch) to gather evidence for and against each hypothesis.
  • Targeted requirements interview: Feeds trace outputs into a deep-interview stage to crystallize the spec with ambiguity reduction instead of starting requirements work from scratch.
  • 3-point injection bridge: Injects a defensible trace-derived root-cause context, substitutes brownfield exploration with trace synthesis, and seeds the first interview questions using per-lane critical unknowns.

Quick Start

Use the deep-dive skill with your problem statement by saying: /deep-dive "Why did my system fail and what should we do next?"

Frequently Asked Questions about deep-dive

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

FAQPage Schema
How do I investigate root causes of ambiguous system failures before planning changes?

Ambiguous system failure investigation requires evidence-grounded causal explanation. A trace-to-interview pipeline runs parallel investigations across code paths, environment configs, and measurement mismatches to isolate root causes before generating requirements.

Can I generate requirements specifications directly from debugging trace outputs?

Requirements generation from debugging traces injects root-cause context into a deep-interview stage. This substitutes manual brownfield exploration with trace synthesis, seeding interview questions using per-lane critical unknowns to crystallize precise specifications.

What's the best way to trace multi-agent system behavior across different execution lanes?

Multi-agent tracing uses a 3-lane causal execution model. It runs parallel investigations for implementation paths, configuration orchestration, and artifact mismatches simultaneously, gathering evidence for and against each hypothesis to resolve behavioral uncertainty.

Does autonomous planning work for resolving uncertainty in complex bug investigations?

Autonomous planning resolves bug investigation uncertainty through stateful orchestration. It persists trace context across stages, allowing interactive lane confirmation and resume behavior when investigating intermittent or complex system failures.

How do I crystallize requirements when the root cause of a bug is still unclear?

Requirements crystallization for unclear root causes uses a 3-point injection bridge. It feeds defensible trace-derived context into targeted interviews, reducing ambiguity by identifying critical unknowns per investigation lane rather than starting from scratch.

When should I use causal investigation instead of standard debugging for system behavior exploration?

Causal investigation suits system behavior exploration where root cause uncertainty blocks planning. When bugs are ambiguous and require parallel evidence gathering across code, config, and measurement layers before any changes can be specified.