| name | enterprise-ai |
| version | 0.1.0 |
| description | Antigravity-native suite of consulting, PM, strategy, prioritization, deck,
and meeting-prep workflows. Use `/enterprise-ai` to route ambiguous business
requests, or call a specific command such as `/decision-memo`, `/storyline`,
`/prioritize`, `/mckinsey-critic`, `/data-insights`, or `/gamma-deck`.
|
| allowed-tools | ["Read","Bash","AskUserQuestion"] |
Enterprise AI
Use this suite when the user needs structured business thinking, executive communication, roadmap prioritization, meeting prep, deck narratives, document critique, or lightweight data/deck automation.
Commands
| Command | Use when the user needs |
|---|
/scpr | Situation, Complication, Problem, Recommendation structure |
/issue-tree | MECE problem decomposition and hypotheses |
/decision-memo | A 1-page yes/no decision memo |
/storyline | Claim-based slide titles and deck flow |
/prioritize | RICE, impact/effort, value/complexity, or weighted scoring |
/meeting-prep | Pre-read, agenda, talking points, objections, rebuttals |
/ai-use-case-score | Personal AI use-case scoring on Value x Feasibility x Safety |
/mckinsey-critic | Consulting-style review of a deck, memo, or strategy |
/deck-pipeline | Strategist -> builder -> critic -> fixer deck workflow |
/data-insights | CSV/Excel analysis with plain-English insights |
/gamma-deck | Gamma API deck generation from a storyline |
Routing
If the user invokes /enterprise-ai, choose the smallest relevant command:
- "Structure this argument" ->
/scpr
- "Break down this problem" ->
/issue-tree
- "Help me get approval" ->
/decision-memo or /meeting-prep
- "Make this into slides" ->
/storyline, then /deck-pipeline if they want a fuller draft/review loop
- "Rank these features" ->
/prioritize
- "What should I AI-ify?" ->
/ai-use-case-score
- "Review this before I send it" ->
/mckinsey-critic
- "Analyze this file" ->
/data-insights
- "Generate a PPTX" ->
/gamma-deck
Token Discipline
Read only the specific sub-skill SKILL.md needed for the user's request. Do not load full references unless the sub-skill says to, the user asks for explanation, or output quality depends on an example/rubric.
Guardrails
- Ask for missing decision-critical inputs before drafting.
- Keep outputs decision-oriented and concise.
- Make assumptions explicit when data is missing.
- Do not invent sources or benchmarks; label estimates as estimates.
