🎙️ Corporate Jargon Translator

An Agentic AI Skill to decode corporate speak, HR euphemisms, executive deflection, and startup IPO hype into raw, unfiltered truth (and vice versa).

Inspired by the viral Corporate Translator video, Corporate Jargon Translator transforms AI assistants into sharp, witty, and deeply accurate corporate decoders. It helps employees see through managerial rhetoric, spot HR traps, understand startup equity realities, and protect their careers.

• Translates emails, manager DMs, town hall notes, and PIP warnings into plain subtext.
• Exposes Startup IPO Hype & "SME" Delusions (e.g. trading "monopoly-money" paper equity for 80-hour work weeks).
• Every decode follows the same three-part output contract:
  1. 🎙️ Inline Translator Overlay — the loaded lines quoted one by one, each with a short deadpan verdict from the Unfiltered Translator. He waits until the jargon is clear; lines with nothing to decode get silence, not speculation.
  2. Summary Decoder Table — every flagged phrase with its subtext and a Risk Level:
     • 🟢 LOW: Harmless office filler.
     • 🟡 MEDIUM: Subtle pushback or scope creep.
     • 🟠 HIGH: Heavy workload dump, title stall, or compensation dilution.
     • 🔴 CRITICAL: PIP, disciplinary track, brain drain, or liquidation risk.
  3. Tactical Employee Advice — 2-3 genuine, actionable survival steps for the receiving party.

Example overlay:

> "Our corporate governance committee has placed a temporary hold on senior title calibrations."
🎙️ Translator: Promotion denied.

> "...the uncompromised spotlight it truly deserves when the capital restructuring is finalized."
🎙️ Translator: "Later" means never.

• Converts raw, angry, or direct employee feedback into polished, executive-ready corporate phrasing.
• Output: Blunt Thought → Corporate HR-Safe Version → Strategic Intent (why the phrasing protects you politically).

Summoned when you ask — "how would all the personas respond to this email?", a single-character ask ("what would Dave say?"), the full show (decode contract first, then the panel), or directly via the /lala-panel companion skill. A comedic, multi-persona panel roast featuring classic workplace archetypes — and the key gag: the Founder and Lala Ji are two different people (the Founder never orders, only celebrates; Lala Ji does the grinding and thinks he's the real CEO):

• 🧘 The Founder ("The Enlightened Visionary"): "We are a FAMILY! Huge shoutout to the 2 AM commits — THIS is ownership!" — never orders, only celebrates.
• 👑 Lala Ji ("The Self-Proclaimed CEO"): "Screw the Founder — I am the real CEO here! Hourly updates. Yes, on Sunday." — the summoned micromanager.
• 👔 The Sycophant Senior ("Chief Defense Officer"): "I deeply appreciate your candor..."
• ☕ The Tired Technical Anchor ("Dave"): "I built the monolith 4 years ago. I'm off-call at 5:00 PM."
• 🍕 The Naive New Joiner ("The True Believer"): "Yay! Pizza party! Our 0.01% equity is going to IPO!"
• 🎙️ The Unfiltered Translator ("The Cynical Realist"): "They don't like you. You're in trouble."

This repository is pre-built to run natively across all major AI agent harnesses:

Default to a harness — it runs the skill on the subscription you already pay for (Claude, ChatGPT/Codex, etc.). The raw API adapter is for embedding the translator in your own scripts and bills per token on top.

Both skills ship together: corporate-jargon-translator (decode/encode) and lala-panel (character panel) live side by side in .agents/skills/ and are discovered by every harness above the same way.

Clone or copy this repository into your project or .agents/skills/ directory. Then prompt your AI agent:

Decode this email from my manager using corporate-jargon-translator:
"We are taking a disciplined approach to cash flow and operational bandwidth..."

In harnesses with slash-command skill invocation (Claude Code, Codex CLI), the skills are directly callable:

/corporate-jargon-translator decode this email: ...
/lala-panel how would everyone react to this memo: ...

Test the pattern matcher and benchmark test suite locally:

# Run automated benchmark test suite
python3 .agents/skills/corporate-jargon-translator/scripts/translate.py --test

# Decode custom text via CLI
python3 .agents/skills/corporate-jargon-translator/scripts/translate.py --decode "We are asking everyone to wear multiple hats."

Only needed outside a harness (e.g. embedding the translator in your own scripts). Generate the API payload:

python3 adapters/codex/openai_api_demo.py

corporate-jargon-translator/
├── README.md                                     # Main project documentation
├── GETTING-STARTED.md                            # Benchmark test prompts & execution guide
├── CLAUDE.md                                     # Repository guide for Claude Code
├── AGENTS.md                                     # Repository guide for Codex CLI & agentskills.io harnesses
├── .claude/
│   ├── rules/
│   │   └── corporate-jargon-translator.md        # Native Claude Code rule file
│   └── skills/
│       ├── corporate-jargon-translator/          # Symlink → .agents/skills/... (Claude Code skill discovery)
│       └── lala-panel/                           # Symlink → .agents/skills/lala-panel
├── .agents/
│   └── skills/
│       ├── lala-panel/
│       │   └── SKILL.md                          # /lala-panel — character panel companion skill
│       └── corporate-jargon-translator/
│           ├── SKILL.md                          # Master Antigravity / OpenCode Skill Spec
│           ├── references/
│           │   ├── jargon_dictionary.md          # Comprehensive subtext & risk lexicon
│           │   ├── character_cast.md             # Lala Company persona definitions
│           │   └── portability_guide.md          # Multi-harness export instructions
│           ├── examples/
│           │   └── video_transcript_demo.md      # Benchmark video transcript analysis
│           └── scripts/
│               └── translate.py                  # Offline CLI matcher & test suite
└── adapters/
    └── codex/                                    # Optional raw-API route (harness route needs none of this)
        ├── system_prompt.txt                     # Copy-pasteable system prompt (source of truth)
        ├── system_prompt.json                    # Generated API payload (openai_api_demo.py --write-json)
        └── openai_api_demo.py                    # Python API payload generator

• Original Video Inspiration: Massive credit to the creator of the viral Corporate Translator video (Manager vs. Employee vs. HR vs. Translator) which sparked the concept for this agentic AI skill.
• Lala Company Archetypes: Inspired by real-world startup workplace dynamics and founder culture.

MIT License. Free for all employees, developers, and AI agents looking to translate corporate speak into truth!