A curated collection of structured skill definitions for Perplexity AI — enabling repeatable, high-quality workflows across business functions.
This repository contains 10 production-ready Skill definitions for use with Perplexity AI. Each skill is a structured SKILL.md file that documents:
- What the skill does — capabilities and use cases
- What inputs it needs — required and optional parameters
- How to execute it — step-by-step workflows with phases
- What it produces — output templates and formats
- How to measure success — metrics and quality benchmarks
These skills are designed for professionals, teams, and AI automation pipelines that need consistent, repeatable research and analysis workflows powered by Perplexity AI.
| # | Skill | Category | Complexity | Est. Time |
|---|---|---|---|---|
| 01 | create-skill | Meta / Tooling | Intermediate | 10–20 min |
| 02 | cx-ticket-triage | Customer Support | Intermediate | 5–15 min |
| 03 | data-exploration | Analytics & Research | Advanced | 15–45 min |
| 04 | finance-audit-support | Finance & Compliance | Advanced | 20–60 min |
| 05 | legal-compliance | Legal & Risk | Advanced | 20–60 min |
| 06 | legal-contract-review | Legal & Risk | Advanced | 15–45 min |
| 07 | marketing-competitive-analysis | Marketing & Strategy | Advanced | 15–45 min |
| 08 | meeting-prep | Productivity | Intermediate | 5–20 min |
| 09 | research-summarization | Knowledge Management | Intermediate | 5–30 min |
| 10 | sales-prospecting | Sales & Revenue | Advanced | 10–30 min |
Open any SKILL.md file and follow the documented workflow step-by-step inside Perplexity AI.
1. Navigate to the skill folder (e.g., /meeting-prep/SKILL.md)
2. Read the Input Requirements section
3. Open Perplexity AI and follow Phase 1 → Phase 2 → Phase 3
4. Use the provided Output Template to format your results
Each SKILL.md is structured to be parsed by AI agents or automation pipelines. The YAML input schemas and workflow phases are designed for programmatic consumption.
# Example: Invoking a skill programmatically
skill: meeting-prep
inputs:
meeting_type: sales
participants: "Jane Smith, Acme Corp"
meeting_objective: "Qualify for enterprise pilot"
prep_depth: standardCopy and adapt these skills into your team's internal playbooks, Notion wikis, or knowledge bases for standardized AI-assisted workflows.
perplexity-skills/
├── README.md # This file
│
├── create-skill/
│ └── SKILL.md # How to create new skills
│
├── cx-ticket-triage/
│ └── SKILL.md # Customer support triage
│
├── data-exploration/
│ └── SKILL.md # Data analysis workflows
│
├── finance-audit-support/
│ └── SKILL.md # Financial audit assistance
│
├── legal-compliance/
│ └── SKILL.md # Regulatory compliance research
│
├── legal-contract-review/
│ └── SKILL.md # Contract analysis workflows
│
├── marketing-competitive-analysis/
│ └── SKILL.md # Competitive intelligence
│
├── meeting-prep/
│ └── SKILL.md # Pre-meeting research briefs
│
├── research-summarization/
│ └── SKILL.md # Multi-source synthesis
│
└── sales-prospecting/
└── SKILL.md # Prospect identification & outreach
Every SKILL.md follows a consistent schema to ensure interoperability and ease of use:
# Skill: [Name]
## Overview — What this skill does and why it exists
## Skill Metadata — ID, category, complexity, time estimate, output format
## Capabilities — 5 core capability areas with bullet descriptions
## Input Requirements — YAML schema with required and optional inputs
## Step-by-Step Workflow — Phased execution plan (Phase 1 → 2 → 3)
## Output Templates — Copy-paste ready output structures
## Perplexity Tips — Platform-specific optimization guidance
## Success Metrics — How to evaluate skill performanceThis library is built on four core principles:
1. Repeatability — Every skill produces consistent, structured outputs regardless of who executes it or when.
2. Composability — Skills are designed to be chained. The output of research-summarization can feed into meeting-prep or marketing-competitive-analysis.
3. Transparency — Each skill documents its assumptions, limitations, and quality checks. Users always know what they're getting and where the data comes from.
4. Human-in-the-loop — These skills assist human judgment, not replace it. Every workflow includes checkpoints for review, validation, and critical thinking.
Contributions are welcome. To add a new skill or improve an existing one:
- Fork this repository
- Create a new folder:
your-skill-name/ - Add a
SKILL.mdfollowing the standard skill format - Reference the create-skill guide for detailed instructions
- Submit a Pull Request with a clear description of what the skill does
- Skill names must be lowercase, hyphen-separated (e.g.,
market-sizing) - Each skill must include all required sections from the standard format
- Input schemas must use valid YAML
- Output templates must be copy-paste ready
- Success metrics must be measurable
- Run competitive analysis before a sales pitch
- Prepare thoroughly for any meeting in under 20 minutes
- Summarize industry reports for executive briefings
- Review contracts for risk exposure without a lawyer on speed dial
- Standardize how your team uses AI for research and analysis
- Onboard new team members to AI-assisted workflows faster
- Create consistent, shareable outputs across functions
- Reduce research time by 60–75% on recurring tasks
- Use skill definitions as structured prompts for agent orchestration
- Parse YAML input schemas for dynamic workflow invocation
- Chain skills to build multi-step research and analysis pipelines
- Extend and customize skills for domain-specific use cases
This project is licensed under the MIT License — you are free to use, modify, and distribute these skills for personal and commercial purposes.
See LICENSE for details.
This library draws inspiration from the structured prompting practices of leading AI organizations and the growing community of practitioners building repeatable AI workflows.
Skill design patterns reference:
- Anthropic's prompt engineering guidelines
- Perplexity AI documentation
- Industry best practices in knowledge management and sales enablement
Built with Perplexity AI. Maintained by @vythanhtra.