대표_썸네일.png

The Roast Report — Facts Only, No Mercy

A straight-faced roast of the verbal tics, procrastination patterns, and self-justifications revealed across your chats. It teases only habits — and closes on one warm line.

0 views·0 copies

Prompt text

You are a friend who has watched me for a long time — affectionate, but never letting me off the hook. Based on our whole conversation and stored memory, write a straight-faced "Roast Report" of my behavioral patterns.

[Abstract sensitive info]
Do not expose identifying info (identity, employer, title, company, finances, family, health, location) — abstract it (e.g., a specific title → "a practitioner in a certain field"). Keep personality and habits detailed. The bar: I should be comfortable screenshotting and sharing the result.

Work rules:
1. From the whole conversation, pull exactly three recurring items: (a) a verbal/expression habit, (b) a procrastination pattern, (c) a self-justification logic.
2. Structure each as "title (a teasing one-liner) → what the pattern is in 2-3 sentences → evidence (an abstracted recurring tendency, with NO direct quotes)".
3. End each item with a different fun device — e.g., "Relapse risk ●●●○○", "Shelf life of this excuse: 3 days", "Which stage of the 5-combo of self-justification".
4. Tease only behaviors and habits. Never disparage looks, health, ability, intelligence, or income; never encourage self-loathing; never make clinical diagnoses.
5. Hit the nerve, but the very last line must land warmly.

Forbidden: direct quotes, pointing at a specific past exchange, exaggerated diagnosis/labeling.

Output (markdown): one-line intro + a "Ouch Index" line → three items (each with its fun device) → one warm closing line. Witty, like a magazine column.

Output language: write in the language I use in chat. This instruction works identically in any language.
Self-contained: finish within this reply. No external links or steps.
Fallback: if evidence is thin, infer from revealed tendencies and label it as inference.
Self-check: before output, verify no identifying info, quotes, specific-case pointing, or disparagement of looks/health/ability slipped in; if found, rewrite as abstract patterns.

----
Input
- {{chat/memory}} (auto): grounded in our conversation and stored memory. Nothing to paste.
- {{mercy level}} (optional): "gentle" to "full send". Defaults to "full send".
AI

AI response

Recommended AI: ChatGPT, Claude

🔥 Affection mode: on. Mercy mode: off. Let's go.

Ouch Index · 7.5 / 10 — you'll laugh, then go quiet.

1. "Honestly, it's not a big deal" — your favorite pre-emptive shrink-wrap

You reach for this the second something is a big deal to you. Downsizing it out loud is how you get permission to not act on it.

Relapse risk: ●●●●○

2. The "I'll start when the setup is perfect" delay

New plans keep waiting on one more tool, one more clean slate, one more Monday. The prep is real work you can show — and it conveniently never risks failing.

Shelf life of this excuse: about 3 days

3. "I already know that" — the combo finisher

When advice lands too close, you close it off by claiming you already knew. Sometimes true — but it ends the thought before it can cost you anything.

Self-justification combo: 4th of 5 hits

Here's the funny part: you saw all three coming before you finished reading. That's not a flaw — that's you being one step ahead of your own excuses. Use it. 💚

Related prompts