skills.rest
theam
theamOFFICIAL
Explore all skills by this author →
Website𝕏@theagilemonkeysGitHub

research-devil-advocate

Stress-test Limina research directions with adversarial evidence and baseline audits.

36|3|Updated Mar 8, 2026
Search Tags:#knowledge base#adversarial analysis#evidence review#research audit#decision checkpoint#experiment critique
One-click install
npx skills add https://github.com/theam/limina --skill research-devil-advocate
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: research-devil-advocate
Source: https://github.com/theam/limina/tree/main/skills/research-devil-advocate
Command: npx skills add https://github.com/theam/limina --skill research-devil-advocate

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you avoid wasting time by adversarially reviewing Limina’s current research direction and next-step plan, focusing on decision quality and evidence traceability rather than looking for reasons to proceed.

Core Features & Use Cases

  • Adversarial research audit: challenges the plan to decide whether it deserves more time, pivoting when evidence or framing is weak.
  • Evidence-grounded critique: reconstructs a claim ledger separating observations, inferences, and recommendations, and treats missing artifacts as a real risk.
  • Decision checkpoint output: produces exactly one of CONTINUE, CONTINUE_WITH_FIXES, PIVOT, STOP, or ESCALATE with an explicit confidence level and the smallest next action that could resolve uncertainty.
  • Audit persistence to the KB: when kb/ exists, writes the review as a CR note and validates KB integrity for reproducibility.

Quick Start

Use research-devil-advocate to review the current direction and recommend whether to CONTINUE, CONTINUE_WITH_FIXES, PIVOT, STOP, or ESCALATE based on the available evidence and traceability.

Frequently Asked Questions about research-devil-advocate

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

FAQPage Schema
How do I run an adversarial research audit to check if my current direction is valid?▼

An adversarial research audit stress-tests your evidence, baselines, and assumptions to determine if your current research direction is decision-relevant. It applies rubric-based critique to challenge whether your plan deserves more time or needs a pivot.

What is a decision checkpoint in research and when do I need one?▼

A decision checkpoint is a step-back review used during pre-experiment, post-experiment, plateau, or pre-commitment phases. It evaluates whether your research direction remains decision-relevant based on available evidence and traceability before committing further resources.

How do I review research evidence and separate observations from inferences?▼

Evidence review reconstructs a claim ledger that explicitly separates observations, inferences, and recommendations. Missing artifacts are treated as real risks, ensuring your critique is grounded in traceable evidence from your knowledge base rather than assumptions.

Can I get a clear status decision like continue or pivot after reviewing my research plan?▼

Yes, the review outputs exactly one status decision: CONTINUE, CONTINUE_WITH_FIXES, PIVOT, STOP, or ESCALATE. Each includes an explicit confidence level and identifies the smallest next experiment or action that could resolve remaining uncertainty.

Does the research audit require an existing knowledge base to function?▼

The audit requires evidence re-grounding from your Limina knowledge base, including mission, active status, and linked artifacts. When a knowledge base directory exists, it persists the review as a CR note and validates KB integrity for reproducibility.

When should I not use an adversarial analysis approach for my research?▼

Adversarial analysis is not suited for routine progress tracking or confirmation-seeking tasks. It is designed for critical decision checkpoints where the next action depends on whether the direction is decision-relevant and requires rigorous evidence traceability.

Related Skills

Explore More Education & Research Skills→
theamtheam
36

limina

Convert measurable problems into evidence-backed research plans with a persistent H->E->F evidence loop.

Official
Intermediate
00PrabalK0000PrabalK00
1

experiment-analysis-assistant

Analyzes A/B test outcomes and proposes likely root causes and next steps.

Community
Intermediate
zbyzby
82

cp-skill-validate

Validate frontmatter, schema, links, and required sections in knowledge base notes.

Community
Advanced
BorisserzBorisserz
1

doubt-driven-development

Runs adversarial fresh-context reviews of non-trivial outputs using doubt cycle CLAIM-EXTRACT-DOUBT-RECONCILE-STOP.

Community
Advanced