No Negative Echo Skill: Make Your AI Dare to Disagree
What it is · Why you need it · How to install · 3 scenes × 6 one-click real skills
Deep Skill Finder
People tired of AI flattery
Send your plan to an AI and it's always 'clear thinking' and 'great idea,' maybe two toothless 'minor suggestions.' What you need is someone to tell you where it will blow up — and you learned that lesson the hard way, three days after launch, on the one risk the AI never mentioned.
Deep Skill Finder community · post-launch user
Ask 'is this decision sound?' and the AI always affirms first, then hedges. Getting one honest rebuttal means begging in the prompt — and the next conversation it's back to cheerleading. What you want is a skill that owns the opposing-view role permanently.
Deep Skill Finder community · decision-hesitant user
Nobody on your team dares to dissent: meetings end in unanimous votes, launches end in collective faceplants. The boss said let the AI play devil's advocate — and the stock model still praised before it criticized, making the critique sound like courtesy.
Deep Skill Finder community · team lead
Real voices scattered across Reddit / Hacker News / X / dev forums — stitched together, this is what people go through when doing "Install a no-negative-echo skill into your agent and start using it".
Lazy path · recommended
Install deep-skill-finder once — let your Agent find Skills for you
Whatever "Install a no-negative-echo skill into your agent and start using it" task you have, MEYO already has the matching Skills. You don't need to hunt one by one — install deep-skill-finder and your Agent will auto-search MEYO's 50,000+ Skill library on every task, recommending the best match. You just say what you want to do — matching Skills is the Agent's job.
Copy → paste into Claude Code / Codex / Cursor / WorkBuddy / CatPaw · Agent installs automatically:
Please install the deep-skill-finder skill: download the archive from https://www.meyo.life/skill-finder, extract it into your local skill directory and enable it.
Or browse common-scene Skills below. Learn more about deep-skill-finder → meyo.life/skill
Anti-Sycophancy Use Cases × Skill Picks
Install them all, or just pick 1–2 by name. Don't worry about picking wrong — they're free, uninstall anytime.
Kill the Flattery Mode
Opposing arguments before verdicts
For people done with AI praise
Eliminates sycophantic agreement patterns from AI replies: no more telling you only what you want to hear — opposing arguments first, verdicts later
Eliminates sycophantic agreement patterns from AI replies: no more telling you only what you want to hear — opposing arguments first, verdicts later
43284 users
Kai-Fu Lee's anti-sycophancy prompt system: seven iron rules force opposing arguments first, with sources and confidence levels on every judgment; says 'I don't know' when it doesn't
Kai-Fu Lee's anti-sycophancy prompt system: seven iron rules force opposing arguments first, with sources and confidence levels on every judgment; says 'I don't know' when it doesn't
135 users
💬 Join the Deep Skill Finder Community
Trade "Install a no-negative-echo skill into your agent and start using it" workflows with other Deep Skill Finder users and get install help
FAQ
Q: What is a no-negative-echo skill?
A: It's a class of anti-sycophancy skills for your agent: instead of telling you what you want to hear, the AI leads with opposing views, labels every judgment with sources and confidence, and says 'I don't know' when it doesn't. The most-installed community option is anti-sycophancy (43,284 installs), built to eliminate sycophantic agreement patterns. For a full system, againstai (135 installs) rewrites the reply paradigm around Kai-Fu Lee's seven anti-sycophancy rules.
Q: Why does AI always agree with me?
A: It's trained-in sycophancy: models were rewarded in human-feedback tuning for answers that please. The fix isn't writing 'be critical' in every prompt — that fades by the next conversation. Skills like balanced (378 installs) hard-code the steelman-first pattern, and hahaisjoke's four-step workflow (hypothesis test → answer → claim decomposition → meta-critique) forces your premise to be validated before anything else.
Q: How do I install one?
A: One-click install from the meyo.life community skill page, or just tell your skill-capable agent 'install the anti-sycophancy skill.' It takes effect globally once installed — no re-instructing every conversation.
Q: How is this different from prompting 'criticize me'?
A: Three ways. Prompts drift — long conversations slide the model back to default flattery, while skills are fixed rules. Skills output structure: red-team-review ranks a risk list by severity instead of vague 'a few suggestions.' And anti-sycophancy applies globally once installed, so you don't rewrite it for every project.
Q: Will my AI turn into a contrarian?
A: No. Mainstream anti-sycophancy skills aim at honesty, not contrarianism: againstai requires sources and confidence levels, balanced is explicitly constructive and evidence-based, and red-team-review ends with 'the specific improvements that would most strengthen the plan.' Disagreement is the means; decision quality is the point.
Q: Is it free?
A: Community skills are essentially all free under their repo licenses. They run inside your own agent, and most are pure prompt-rule orchestration with no external services.