Cline Skills
https://github.com/cline/skills- Category
- AI Agents
- Rank
- No. 1540Tools index
- Pricing
- Open Source
- Type
- AGENT
- Builder
- cline
- GitHub
- 23 stars
- Date
About
Official collection of skills used at Cline, the open-source AI coding agent for VS Code.
What it does
Cline Skills is a portable library of task playbooks for coding agents. An agent selects a relevant skill from conversation context, then loads its instructions, references, and sometimes bundled scripts for specialized work such as database operations, research, code review, documentation, and session analysis.
Why it's ranked here
The collection is worth considering because it combines broad subject coverage with concrete operating procedures. Several skills include safety rules, staged workflows, reference routing, and executable helpers. Its Agent Skills compatibility also avoids locking the material to one host, although usefulness varies by task and required external service.
What's good
The strongest skills explain decisions, not merely commands. The Aurora DSQL material separates schema, access control, migrations, query plans, and troubleshooting into targeted references. Session tooling accounts for duplicated transcript entries and split API responses. Installation options cover marketplaces, a general skills CLI, and manual copying.
Tradeoffs
This is a heterogeneous collection rather than one consistent application. Some entries provide detailed references and scripts, while others are concise operational guides. Several workflows depend on external products, credentials, CLIs, MCP servers, or service access. Manual cloning also requires submodules to obtain the SDK and ClickHouse material.
How to use it well
Use it when your agent supports the Agent Skills standard and you repeatedly handle one of its covered domains. Install the collection, then let contextual triggers select focused guidance. It suits agent-assisted engineering and operations workflows. It does not replace the external databases, cloud accounts, credentials, CLIs, or MCP servers those workflows require.
Technical notes+
README.md defines contextual loading, installation surfaces, supported agents, and the repository layout. Each skill follows the SKILL.md convention with name and description frontmatter plus optional references and scripts. .claude-plugin/plugin.json, .claude-plugin/marketplace.json, .cursor-plugin/plugin.json, and .cursor-plugin/marketplace.json expose marketplace packaging. skills/cline-sdk and multiple ClickHouse entries under skills/data-analyst/skills/ are symlinks backed by git submodules in .vendor/. skills/session-report/analyze-sessions.mjs is a Node CLI that streams JSONL transcripts, deduplicates entries by UUID and API usage by request ID, and supports text or JSON output. skills/cline-session-search/scripts/search_sessions.py is a Python CLI that searches JSON metadata and a read-only SQLite session database.
Observed
- License
- Apache License 2.0
- Content format
- Agent Skills use Markdown instructions with YAML frontmatter and optional reference files.
- Implementation languages
- The supplied repository text includes JavaScript and Python command-line scripts.
- Installation
- Install through the npx skills CLI, Claude Code marketplace, Cursor Marketplace, or manual clone and copy.
- Agent compatibility
- Documented support includes Cline, Claude Code, Cursor, OpenCode, OpenAI Codex, and Pi.
- Interfaces
- The collection includes agent instructions, bundled CLI scripts, and skills that guide MCP-based workflows.
- Repository structure
- The Cline SDK and official ClickHouse skills are supplied through symlinks backed by git submodules.
Read from README.md, skills/session-report/analyze-sessions.mjs, skills/cline-session-search/scripts/search_sessions.py, LICENSE, .claude-plugin/plugin.json, .cursor-plugin/plugin.json, .claude-plugin/marketplace.json, .cursor-plugin/marketplace.json, skills/dsql/SKILL.md, skills/mintlify/SKILL.md, skills/oracledb/SKILL.md, skills/dr-bedrock/SKILL.md, skills/exa-search/SKILL.md.
What it can do
Generate code from natural language prompts
Natural language description of desired functionality → Source code in various programming languages
Debug and fix code issues
Source code with bugs or errors → Corrected code with fixes and explanations
Refactor existing code
Existing source code → Improved, optimized, or restructured code
Explain code functionality
Source code snippets or files → Natural language explanations of code behavior
Create unit tests
Source code functions or modules → Test cases and test code
Review code quality
Source code files → Code quality assessment and improvement suggestions
Generate documentation
Source code and project structure → Technical documentation and code comments
Tags
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.