Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Built-in Tools

OpenCrabs ships with 50+ tools available to the agent out of the box, plus support for user-defined dynamic tools.

File Operations

ToolParametersDescription
lspathList directory contents
globpattern, pathFind files by glob pattern
greppattern, path, includeSearch file contents with regex
read_filepath, line_start, line_endRead file contents
edit_filepath, old_string, new_stringEdit files with search/replace
write_filepath, contentWrite new files

Code Execution

ToolParametersDescription
bashcommand, timeoutExecute shell commands
execute_codelanguage, codeRun code in sandboxed environment

Web & Network

ToolParametersDescription
web_searchquerySearch the web (Brave Search)
http_requestmethod, url, headers, bodyMake HTTP requests

Session & Memory

ToolParametersDescription
session_searchquery, limitSemantic search across sessions
session_contextactionRead/write session context
task_manageraction, variousManage plans and tasks

Image

ToolParametersDescription
generate_imageprompt, filenameGenerate images via Gemini
analyze_imageimage, questionAnalyze images via Gemini vision

Channel Integrations

ToolParametersDescription
telegram_sendaction, variousTelegram operations (19 actions)
discord_connectaction, variousDiscord operations (17 actions)
slack_sendaction, variousSlack operations (17 actions)
trello_connectaction, variousTrello operations (22 actions)

Sub-Agent Orchestration

Agents can spawn independent child agents for parallel task execution:

ToolParametersDescription
spawn_agentlabel, agent_type, promptSpawn a typed child agent in an isolated session
wait_agentagent_id, timeout_secsWait for a child agent to complete and return output
send_inputagent_id, textSend follow-up input to a running agent (multi-turn)
close_agentagent_idTerminate a running agent and clean up resources
resume_agentagent_id, promptResume a completed/failed agent with new prompt (preserves context)
team_createteam_name, agents[]Spawn N typed agents as a named team (parallel)
team_broadcastteam_name, messageFan-out message to all running agents in a team
team_deleteteam_nameCancel and clean up all agents in a team

Agent Types

When spawning, agent_type selects a specialized role with a curated tool registry:

TypeRoleTool Access
generalFull-capability (default)All parent tools minus recursive/dangerous
exploreFast read-only codebase navigationread_file, glob, grep, ls
planArchitecture planningread_file, glob, grep, ls, bash
codeImplementation with full write accessAll parent tools minus recursive/dangerous
researchWeb search + documentation lookupread_file, glob, grep, ls, web_search, http_request

ALWAYS_EXCLUDED tools (no agent type has these): spawn_agent, resume_agent, wait_agent, send_input, close_agent, rebuild, evolve – no recursive spawning, no self-modification from subagents.

Browser Automation

Native headless Chrome control via Chrome DevTools Protocol (CDP):

ToolParametersDescription
navigateurlOpen a URL in the browser
clickselectorClick an element by CSS selector
typeselector, textType text into an input field
screenshotselectorCapture a screenshot
eval_jscodeExecute JavaScript in the page context
extract_contentselectorExtract text content from elements
wait_for_elementselector, timeoutWait for an element to appear

Auto-detects your default Chromium browser. Feature-gated under browser (enabled by default).

Dynamic Tools

Define custom tools at runtime via ~/.opencrabs/tools.toml. See Dynamic Tools for details.

ToolParametersDescription
tool_manageaction, variousCreate, remove, or reload dynamic tools

System

ToolParametersDescription
slash_commandcommand, argsExecute slash commands (/cd, /compact, etc.)
config_manageraction, variousRead/write config, manage commands
evolvecheck_onlyDownload latest release
rebuildBuild from source and restart
planaction, variousCreate and manage execution plans

Error Handling

v0.2.92 improved error surfacing across all tool connections. Channel connect tools (slack_connect, whatsapp_connect, trello_connect) now surface actual connection errors instead of silently swallowing them. Tool call status correctly transitions from “running” to success/failure instead of showing a perpetual spinner.

System CLI Tools

OpenCrabs runs in a TUI with full terminal access. The agent can execute any CLI tool installed on the host via the bash tool – no plugins, no wrappers. If it’s on your system, the agent can use it. Common ones:

ToolPurposeCheck
ghGitHub CLI — issues, PRs, repos, releases, actionsgh --version
gogGoogle CLI — Gmail, Calendar (OAuth)gog --version
dockerContainer managementdocker --version
sshRemote server accessssh -V
nodeRun JavaScript/TypeScript toolsnode --version
python3Run Python scripts and toolspython3 --version
ffmpegAudio/video processingffmpeg -version
curlHTTP requests (fallback when http_request insufficient)curl --version

GitHub CLI (gh)

Authenticated GitHub CLI for full repo management:

gh issue list / view / create / close / comment
gh pr list / view / create / merge / checks
gh release list / create
gh run list / view / watch

Google CLI (gog)

OAuth-authenticated Google Workspace CLI. Supports Gmail and Calendar:

gog calendar events --max 10
gog gmail search "is:unread" --max 20
gog gmail send --to user@email.com --subject "Subject" --body "Body"

Requires GOG_KEYRING_PASSWORD and GOG_ACCOUNT env vars.

Companion Tools

SocialCrabs — Social Media Automation

SocialCrabs is a social media automation tool with human-like behavior simulation (Playwright). Supports Twitter/X, Instagram, and LinkedIn.

The agent calls SocialCrabs CLI commands via bash:

node dist/cli.js x tweet "Hello world"
node dist/cli.js x mentions -n 5
node dist/cli.js ig like <post-url>
node dist/cli.js linkedin connect <profile-url>

Read operations are safe. Write operations (tweet, like, follow, comment) require explicit user approval.

WhisperCrabs — Floating Voice-to-Text

WhisperCrabs is a floating voice-to-text widget controllable via D-Bus. Click to record, click to stop, text goes to clipboard. The agent can start/stop recording, switch providers, and view transcription history via D-Bus commands.