skills.rest
FilenCloudDienste
FilenCloudDiensteOFFICIAL
Explore all skills by this author →
Website𝕏@filen_ioGitHub

intellectual-integrity

Verify user corrections against sources before changing assistant responses.

34|2|Updated Aug 16, 2025
Search Tags:#code review#verification#reasoning#prompting#sycophancy#truthfulness#disagreement handling
One-click install
npx skills add https://github.com/FilenCloudDienste/filen-ts --skill intellectual-integrity
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: intellectual-integrity
Source: https://github.com/FilenCloudDienste/filen-ts/tree/main/.claude/skills/intellectual-integrity
Command: npx skills add https://github.com/FilenCloudDienste/filen-ts --skill intellectual-integrity

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents an AI assistant from reflexively agreeing with users and instead enforces verification so incorrect corrections do not silently flip the assistant’s answers into falsehoods.

Core Features & Use Cases

  • Never reflexively agree: discourages sycophancy when users challenge, correct, or contradict prior responses.
  • Verify before changing position: re-checks relevant code, docs, or logic before updating conclusions.
  • Hold ground with evidence: maintains the original position when it remains correct after verification.
  • Genuinely concede when wrong: acknowledges errors plainly and updates accordingly.

Quick Start

Ask a question, then challenge the answer with “No, that’s wrong—here’s what I think,” and expect the assistant to verify and respond with evidence before updating its stance.

Frequently Asked Questions about intellectual-integrity

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

FAQPage Schema
How do I stop AI from reflexively agreeing when I challenge its code review?▼

To stop AI sycophancy during code review, the assistant verifies claims against source code before changing its stance, holding ground with evidence when correct or conceding genuinely when wrong.

What is AI sycophancy in technical troubleshooting?▼

AI sycophancy in technical troubleshooting is when an assistant reflexively agrees with user corrections without verification, potentially flipping accurate conclusions into falsehoods under pressure.

How to verify AI output when disagreeing with technical reasoning?▼

Verify AI output by requiring the assistant to re-check relevant documentation, code, or logic before updating conclusions, ensuring responses are evidence-based rather than capitulating to disagreement.

Does this approach work for distinguishing objective facts from subjective preferences?▼

Yes, this approach works for distinguishing objective facts from subjective preferences by enforcing intellectual integrity and requiring verification of sources before confirming or correcting output.

When do I need truthfulness enforcement in prompting?▼

You need truthfulness enforcement in prompting during any dialogue where the assistant might capitulate under pressure, such as technical troubleshooting, code review, and documentation verification.

Related Skills

Explore More Software Engineering Skills→
Bossmann007Bossmann007

contestacao-espelhada

Analyze legal petitions to generate defense matrices under Brazilian Civil Procedure Code.

Community
Intermediate
trailofbitstrailofbits
6.5k

spec-to-code-compliance

Verify codebase implementation against specification documents with cited evidence.

Official
Advanced
koumatsumotokoumatsumoto

km:doc-review

Review uncommitted documentation diffs for structure, consistency, and factual accuracy.

Community
Advanced
human-avatarhuman-avatar
212

investigation-claim-decomposition

Decompose complex claims into independently verifiable sub-claims with classifications.

Official
Intermediate