A collection of Agent Skills we use at Cline and want to share with the community.
These skills work with any agent that supports the Agent Skills standard, including Cline, Claude Code, Cursor, OpenCode, OpenAI Codex, and Pi.
Install using the npx skills CLI:
npx skills add https://github.com/cline/skills
Install using the plugin marketplace:
/plugin marketplace add cline/skills
/plugin install cline@cline
Install from the Cursor Marketplace or add manually via Settings > Rules > Add Rule > Remote Rule (Github) with cline/skills.
Clone this repo (with submodules) and copy the skill folders into the appropriate directory for your agent:
git clone --recurse-submodules https://github.com/cline/skills
| Agent | Skill Directory | Docs |
|---|---|---|
| Cline | ~/.cline/skills/ |
docs |
| Claude Code | ~/.claude/skills/ |
docs |
| Cursor | ~/.cursor/skills/ |
docs |
| OpenCode | ~/.config/opencode/skills/ |
docs |
| OpenAI Codex | ~/.codex/skills/ |
docs |
| Pi | ~/.pi/agent/skills/ |
docs |
Skills are contextual and auto-loaded based on your conversation. When a request matches a skill's triggers, the agent loads and applies the relevant skill to provide accurate, up-to-date guidance.
| Skill | Useful for |
|---|---|
| cline-sdk | Building AI agents with the Cline SDK: Agent runtime, ClineCore sessions, custom tools, plugins, events, LLM providers, scheduling, multi-agent teams, and production deployment |
| review-team | Running a fleet of specialized reviewer subagents (correctness, security, architecture, conventions, simplicity, UX, reliability, telemetry, testing, compatibility, docs) against the same change, single-pass or iterate-until-clean |
| linear-sdk-scripting | Doing Linear work from the terminal without the Linear MCP: list, open, create, update, close, and comment on issues, and query teams, projects, cycles, users, and workflow states by writing small Node scripts against the official @linear/sdk with a personal API key |
| amazon-location-service | Building AWS location-aware apps with maps, places, geocoding, routing, geofences, and trackers using Amazon Location Service |
| amplify-workflow | Building and deploying full-stack web and mobile apps with AWS Amplify Gen2, including auth, data, storage, functions, APIs, and AI Kit patterns |
| attorney-assist | Connecting users with LegalZoom attorney consultation workflows, including plan checks, matter context gathering, topic matching, scheduling, and failure guardrails |
| building-pydantic-ai-agents | Building Pydantic AI agents with tools, capabilities, structured output, streaming, YAML-defined agents, testing, hooks, and multi-agent patterns |
| convex-design | Designing and building reactive, type-safe Convex backends with schemas, queries, mutations, actions, auth, storage, scheduling, real-time features, and agent workflows |
| cosmosdb-best-practices | Designing, reviewing, and optimizing Azure Cosmos DB NoSQL data models, partition keys, queries, SDK usage, throughput, and vector search |
| dataproc-skills | Managing Google Cloud Dataproc clusters and jobs through bundled Node scripts for listing, inspection, job submission, and cancellation |
| desktop-commander-overview | Using Desktop Commander MCP capabilities for persistent shells, long-running processes, broader filesystem access, structured files, search, SSH, and process management |
| dsql | Building with Amazon Aurora DSQL, including schemas, SQL execution, migrations, query plans, IAM auth, ORM migration, bulk loading, MCP tools, and distributed SQL patterns |
| exa-search | Running Exa-powered research workflows for deep dives, lead generation, literature reviews, competitive analysis, and multi-step search synthesis |
| firestore-data | Working with Firestore collections and documents through bundled Node scripts for CRUD, data retrieval, and collection hierarchy exploration |
| frontend-design | Creating distinctive, production-grade frontend interfaces with strong visual direction, refined typography, cohesive color, motion, and polished implementation details |
| gcp-to-aws | Planning and implementing migrations from Google Cloud to AWS, including infrastructure, databases, AI workloads, agent stacks, Terraform, and cutover strategy |
| knowledge-catalog-discovery | Discovering and exploring Knowledge Catalog data assets, entries, entry groups, entry types, and metadata through bundled scripts |
| math-olympiad | Solving and verifying competition math problems with adversarial proof checks, counterexample hunting, calibrated confidence, and optional LaTeX output |
| mintlify | Building Mintlify documentation sites, including docs.json, navigation, MDX pages, components, and API reference configuration |
| oracledb | Managing and monitoring Oracle databases with SQL execution, schema exploration, query plans, session monitoring, resource checks, and object health scripts |
| playground | Building self-contained interactive HTML playgrounds with controls, live previews, presets, and copyable prompts for design, data, code, document, and concept exploration |
| sap-fiori-add-visual-filter | Adding chart-based visual filters to SAP Fiori Elements filter bars and value help dialogs using CAP or ABAP RAP annotations |
| save-to-spotify | Producing polished audio episodes with narration, cover art, timeline companions, show notes, and Save to Spotify CLI workflows |
| searching-sourcegraph | Searching and navigating code with Sourcegraph MCP tools, including keyword search, semantic search, references, definitions, commits, diffs, and Deep Search |
| sentry-cli | Using the Sentry CLI for issues, events, traces, logs, releases, projects, organizations, API calls, authentication, and local Spotlight workflows |
| session-report | Generating self-contained HTML usage reports from Claude Code session transcripts, including token usage, cache behavior, subagent activity, skill usage, and expensive prompts |
| site-specification | Extracting structured website specifications from theme or site descriptions, including site brief, audience, layout notes, brand tone, and typography guidance |
| skill-creator | Creating, improving, evaluating, benchmarking, packaging, and optimizing Agent Skills with bundled schemas, graders, scripts, and review tooling |
| spanner-data | Exploring Spanner database structure, discovering schema objects, and executing SQL through bundled Node scripts |
| teamcity-cli | Driving TeamCity CI/CD from the terminal with the teamcity CLI for builds, logs, jobs, queues, agents, pools, projects, pipelines, and REST API calls |
| ui5-typescript-conversion | Converting SAPUI5/OpenUI5 JavaScript projects to TypeScript while preserving comments, types, XMLView assumptions, tests, and UI5 project conventions |
| use-railway | Operating Railway infrastructure, including signup/login, project and service creation, deployments, variables, domains, logs, metrics, docs, and agent tooling |
| vibe-prospecting | Finding company and contact data, building lead lists, enriching prospects, researching accounts, and chaining prospecting workflows through the Vibe Prospecting CLI |
| endor-setup | Installing and authenticating endorctl, handling namespace selection, and running Endor Labs security scans from non-interactive agent shells |
| windsor-ai-business-data | Querying Windsor.ai business data across marketing, sales, CRM, ecommerce, finance, and analytics connectors for dashboards, reports, and schema exploration |
| data-analyst | Acting as an interactive data analyst over ClickHouse: clarify the actual question first, then connect (local or ClickHouse Cloud) and run safe, bounded SQL via the clickhousectl CLI. Includes a clickhouse sub-skill for CLI auth (browser OAuth) and querying |
skills/
cline-sdk/ -> .vendor/sdk-skill/skill/cline-sdk (git submodule, see below)
review-team/ Multi-reviewer code review fleet
linear-sdk-scripting/ Drive Linear via the @linear/sdk in Node scripts
data-analyst/ Interactive ClickHouse data analyst
skills/clickhouse/ Sub-skill: connect + query via clickhousectl
skills/reading-data-dict/ Sub-skill: resolve metric definitions
skills/steering-user-elicitation/ Sub-skill: clarify before querying
skills/analyzer/ Sub-skill: turn results into findings
skills/plotting/ Sub-skill: charts and visual artifacts
skills/artifact-management/ Sub-skill: save and describe outputs
skills/clickhouse-best-practices/ -> .vendor/clickhouse-agent-skills (official, see below)
skills/chdb-sql/ -> .vendor/clickhouse-agent-skills
skills/chdb-datastore/ -> .vendor/clickhouse-agent-skills
skills/clickhousectl-local-dev/ -> .vendor/clickhouse-agent-skills
skills/clickhousectl-cloud-deploy/ -> .vendor/clickhouse-agent-skills
skills/clickhouse-architecture-advisor/ -> .vendor/clickhouse-agent-skills
skills/clickhouse-js-node-coding/ -> .vendor/clickhouse-agent-skills
skills/clickhouse-js-node-troubleshooting/ -> .vendor/clickhouse-agent-skills
.vendor/
sdk-skill/ Submodule of https://github.com/cline/sdk-skill
clickhouse-agent-skills/ Submodule of https://github.com/ClickHouse/agent-skills (Apache-2.0)
skills/cline-sdk is a symlink into the cline/sdk-skill submodule, which is maintained in its own repo. The official ClickHouse skills under skills/data-analyst/skills/ are symlinks into the ClickHouse/agent-skills submodule (Apache-2.0). To pull the latest:
git submodule update --remote .vendor/sdk-skill
git submodule update --remote .vendor/clickhouse-agent-skills
Open a PR. New skills should follow the Agent Skills spec: a SKILL.md at the skill directory root with a frontmatter name and description, plus any supporting reference files.