Boss Whisperer
b1rdmania/boss-whisperer-skill
Decodifique seu chefe: descubra o que ele realmente quer em cada mensagem
Ver no GitHubUma skill para Claude que decodifica mensagens do seu gestor (emails, Slack, feedback) usando frameworks de Munger e Gottman para revelar o que ele realmente quer, seus gatilhos e a melhor estratégia de resposta. Funciona em dois modos: análise rápida de mensagens individuais ou perfil aprofundado usando fontes públicas.
Instalação
npx convez-ai b1rdmania/boss-whisperer-skill1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Decodificar emails e feedbacks vagos do gestor
- Preparar resposta estratégica para pedidos de promoção ou orçamento
- Entender dinâmicas de poder e gatilhos emocionais do seu chefe
- Analisar padrões em comunicação de fundadores e líderes
Veredito

Ferramenta polêmica mas prática para gerenciamento ascendente em ambientes corporativos confusos. Usa psicologia aplicada (Munger, Gottman) para traduzir comunicação velada — útil para quem precisa navegar política organizacional, mas requer maturidade ética ao aplicar.
README
A tradução para o português deste README ainda está na fila.
Boss Whisperer
Reverse-engineer your manager. CIA-grade profiling for corporate bullshit.
Boss Whisperer is a Claude skill that reads the messages you already have — emails, Slack threads, founder rambles, performance feedback, "quick thoughts" — and gives you the real read:
- what they actually mean
- what they want
- what they are avoiding
- what will trigger them
- what not to say
- the reply most likely to work
It is built for managing up, surviving vague feedback, decoding founder psychology, and turning corporate fog into usable intel.
Don't want to install anything? Run the Lite version free in your browser at bosswhisperer.fun. No signup, nothing stored.
Why this exists
Most workplace advice is HR cosplay.
Boss Whisperer is for the actual game: unclear power, status anxiety, hidden incentives, fragile egos, vague asks, fake urgency, and managers who say "let's circle back" when they mean "I don't want to decide."
Paste the message. Get the read. Write the reply.
What it does
Two modes.
1. The Read
Fast decode from the material you paste. Use it when you have:
- an email you do not trust
- a Slack thread with bad vibes
- performance feedback that feels coded
- a founder monologue
- a promotion conversation
- a manager who keeps moving the goalposts
It gives you a clean operating read: the hidden ask, the emotional logic, the power dynamic, the likely fear, the decision style, the trap, the best response.
2. The Dossier
A deeper profile using public material. Use it when you want to understand the person behind the messages: founder interviews, LinkedIn posts, public talks, blog posts, podcasts, company updates, investor language.
It folds public positioning into the read: stated values vs actual incentives, status markers, repeated language, ego handles, risk posture, decision triggers, likely blind spots, and how to frame the ask.
Public sources only.
Install
git clone https://github.com/b1rdmania/boss-whisperer-skill ~/.claude/skills/boss-whisperer
Then open Claude and say:
decode my boss
Or paste the thread and ask:
what is actually going on here?
Example use cases
My boss sent this after I asked for a promotion. What do they actually mean?
Decode this founder Slack thread. I need budget approval without triggering a defensive spiral.
This feedback feels vague and political. What is the real issue?
Build me a Dossier from these public links, then tell me how to pitch this person.
Output
A structured read:
- The translation — what they are really saying.
- The power dynamic — who has leverage, who wants what, and where the tension sits.
- The emotional logic — what they may be protecting, avoiding, or trying to control.
- The trap — what you should not say.
- The angle — how to frame the ask so it lands.
- The reply — a direct message you can send.
- The next move — what to do if they dodge, delay, or escalate.
Under the hood
Three interpretive lenses:
- Munger inversion — incentives, misjudgment, bias, avoidance, ego protection.
- Therapeutic read — CBT, IFS, Transactional Analysis, attachment patterns, Big Five, DISC.
- Gottman-style communication read — defensiveness, contempt, stonewalling, repair attempts, bids for connection.
The point is not to diagnose anyone. The point is a useful operating model.
Privacy
The local skill runs inside your own Claude. The free browser version processes your message to write the read, then drops it.
- No account
- No server-side storage
- No training dataset
- No boss database
- No messages stored by us
For Dossier mode, you provide the public material or connect your own research tool.
Requirements
For The Read: Claude. (The browser version is free — no key, no signup.)
For The Dossier: a Perplexity MCP server, or any web-search tool your Claude setup exposes, or public links and text pasted manually. Without research access, The Read still works fully.
Run it in the browser
bosswhisperer.fun — the free Lite read. No signup, nothing stored.
Files
SKILL.md— how the skill runsFRAMEWORKS.md— the decode engineDECODE-FORMAT.md— the output format
License
MIT