codeboltai
9 verified skills
.agent/skills/codebolt-provider-development
--- name: codebolt-provider-development description: Create Codebolt providers that connect applications to external environments running remote executors. Providers abstract environment management (Docker, AWS EC2, WSL, SSH, Kubernetes, Git worktree, or any custom environment) and handle all communication with agents. Use when: (1) Building new providers for any environment type, (2) Implementing environment lifecycle management, (3) Setting up remote executor communication, (4) Handling file o
development
codebolt-mcp-access
Core skill for Codebolt agents to interact with the Codebolt MCP (Model Context Protocol) API. Use when agents need to execute tools for file operations, git, browser automation, terminal, memory, orchestration, planning, testing, collaboration, or any Codebolt platform functionality. Triggers on requests involving codebolt.tools.executeTool(), file system operations, git commands, browser control, agent orchestration, task management, or any MCP namespace operations.
tools
skills/codebolt-provider-development
--- name: codebolt-provider-development description: Create Codebolt providers that connect applications to external environments running remote executors. Providers abstract environment management (Docker, AWS EC2, WSL, SSH, Kubernetes, Git worktree, or any custom environment) and handle all communication with agents. Use when: (1) Building new providers for any environment type, (2) Implementing environment lifecycle management, (3) Setting up remote executor communication, (4) Handling file o
development
codebolt-api-access
Use when you need to call codebolt modules (fs, browser, terminal, git, chat, llm, thread, todo, memory, task, swarm, job, roadmap, autoTesting, mcp, actionPlan)
tools
codebolt-agent-development
Build AI agents for the Codebolt platform using @codebolt/agent. Use when creating agents, configuring the agent loop, writing custom message modifiers, implementing processors, creating tools, building workflows, ActionBlocks, or choosing between abstraction levels. Covers Remix Agents (no-code), Level 1 (direct APIs), Level 2 (base components), Level 3 (high-level CodeboltAgent), and ActionBlocks for reusable logic.
tools
codebolt-agent-development
Build AI agents for the Codebolt platform using @codebolt/agent. When creating an agent, first decide what type you need — (1) CodeboltAgent wrapper for simple prompt-based agents with no custom loop, (2) Base Components for customizing the agentic loop, or (3) Core API for full advanced control. Also covers Remix Agents (no-code), ActionBlocks, tools, workflows, and processors. Permissions are handled by the application, not by the agent.
tools
codebolt-agent-development
Build AI agents for the Codebolt platform using @codebolt/agent. Use when creating agents, configuring the agent loop, writing custom message modifiers, implementing processors, creating tools, building workflows, ActionBlocks, or choosing between abstraction levels. Covers Remix Agents (no-code), Level 1 (direct APIs), Level 2 (base components), Level 3 (high-level CodeboltAgent), and ActionBlocks for reusable logic.
tools
codebolt-api-access
Use when you need to call codebolt modules (fs, browser, terminal, git, chat, llm, thread, todo, memory, task, swarm, job, roadmap, autoTesting, mcp, actionPlan)
tools
codebolt-mcp-access
Core skill for Codebolt agents to interact with the Codebolt MCP (Model Context Protocol) API. Use when agents need to execute tools for file operations, git, browser automation, terminal, memory, orchestration, planning, testing, collaboration, or any Codebolt platform functionality. Triggers on requests involving codebolt.tools.executeTool(), file system operations, git commands, browser control, agent orchestration, task management, or any MCP namespace operations.
tools