DeepSeek Skill Leaderboard
Map codebases into navigable knowledge graphs.
Plan, review, test, and ship software with specialist agents.
Reduce agent verbosity and context consumption.
Audit UX and refine frontend design.
Enforce design tokens and accessibility standards.
Find and assemble reusable agent workflows.
Reverse-engineer software and investigate security targets.
Preserve task plans and findings across agent sessions.
Create and publish content across social platforms.
Distill personal knowledge and working styles into agent profiles.
Coordinate research, experiments, and paper writing.
Write precise prompts tailored to AI tools.
Train, evaluate, and research machine learning models.
Edit native office documents alongside an AI assistant.
Turn system descriptions into editable draw.io architecture diagrams.
Create motion design briefs and brandkits.
Remove generic patterns from interfaces, prose, and code.
Coordinate persistent teams of coding agents.
Plan, design, debug, and review software changes.
Turn explanations into readable offline HTML pages.
Generate and edit images and extract visual prompts.
Add image understanding to text-only agents.
Rewrite fiction and professional prose to remove AI patterns.
Coordinate personal tasks across devices with a persistent agent.
Distill an author's writing style into reusable guidelines.
Coordinate coding tasks from implementation to reviewed commits.
Explain code changes with animated architecture diagrams.
Give text-only models image understanding through vision APIs.
Find reusable workflows for development and business tasks.
Track emerging AI names and search opportunities.
Build MCP integrations and inspect coding-agent traffic.
Develop software and automate operations with Antigravity.
Apply structured reasoning to decisions and complex problems.
Review architecture, technical debt, and test quality.
Turn product research into plans and stakeholder updates.
Automate development, research, and content workflows.
Plan coding changes and verify completion with evidence.
Compare model API prices across providers.
Route agent decisions through TypeSafe's Jev model.
Audit websites for AI search discoverability.
Connect phone calls to conversational AI agents.
Benchmark, profile, and troubleshoot LLM serving infrastructure.
Translate videos with synchronized dubbing and subtitles.
Research, write, and review scientific manuscripts.
Create UI designs, presentations, and graphics on canvas.
Search the web and retrieve structured source evidence.
Plan, execute, and analyze multi-brand marketing strategies.
Audit and improve brand visibility in AI search.
Build neuro-symbolic AI agents with Synalinks.
Research literature and develop academic papers and presentations.