OB1

https://www.openblocklabs.com/manual
Admin
OB1 is a multi-model terminal-based coding agent with self-optimizing workflows, supporting 300+...
Tokens:5,978
Snippets:65
Trust Score:7.7
Update:7 months ago
Show doc for...
Tokens:
Raw
### Example Hook Script for Blocking Secrets

Source: https://www.openblocklabs.com/manual

This bash script demonstrates how to write a hook script for OB1. It reads JSON input from stdin, extracts the content being written to a file, and can be extended to check for and block sensitive information.

```bash
#!/usr/bin/env bash
# .ob1/hooks/block-secrets.sh
input=$(cat)

# Extract content being written
content=$(echo "$input" | jq -r '.tool_input.content // ""')


```

--------------------------------

### Configure OB1 Hooks for File Operations

Source: https://www.openblocklabs.com/manual

This JSON configuration sets up hooks to automate tasks after specific OB1 actions. In this example, an 'npm run lint --fix' command is executed after OB1 writes or replaces a file, ensuring code quality.

```json
{
  "hooks": {
    "AfterTool": [
      {
        "matcher": "write_file|replace",
        "hooks": [
          {
            "type": "command",
            "command": "npm run lint --fix"
          }
        ]
      }
    ]
  }
}
```

--------------------------------

### Configure MCP Servers in OB1 Settings

Source: https://www.openblocklabs.com/manual

Manual configuration of MCP servers by editing the `mcpServers` key in the `~/.ob1/settings.json` file. This allows for direct specification of server URLs or commands and arguments for local servers.

```json
{
  "mcpServers": {
    "linear": {
      "url": "https://mcp.linear.app/sse"
    },
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@anthropic-ai/mcp-filesystem", "/path/to/dir"]
    }
  }
}
```

--------------------------------

### Configure Per-Model System Prompts and Budgets

Source: https://www.openblocklabs.com/manual

This JSON configuration allows customization of OB1's behavior for specific models. You can inject custom system prompts, set thinking budgets, and define behavioral overrides. OB1 automatically includes model identity information.

```json
{
  "modelConfigs": {
    "anthropic/claude-opus-4": {
      "systemPromptPrefix": "You are a senior architect...",
      "thinkingBudget": 10000
    }
  }
}
```

--------------------------------

### Add MCP Servers via OB1 CLI

Source: https://www.openblocklabs.com/manual

CLI commands for adding Model Context Protocol (MCP) servers to OB1. Supports both SSE/HTTP transport for remote servers and stdio transport for local servers, allowing extension of OB1 with custom tools.

```bash
# SSE/HTTP transport (remote servers)
ob1 mcp add linear https://mcp.linear.app/sse --transport sse --scope user

# stdio transport (local servers)
ob1 mcp add filesystem npx -- -y @anthropic-ai/mcp-filesystem /path/to/dir --scope user
```

--------------------------------

### Launch Chrome Browser Profiles

Source: https://www.openblocklabs.com/manual

Commands to launch the OB1-managed Chrome browser with persistent session support. Supports default profiles or custom named profiles.

```bash
/chrome               # Launch Chrome with default OB1 profile
/chrome --profile work  # Use a named profile
```

--------------------------------

### Manage Git Worktrees in OB1

Source: https://www.openblocklabs.com/manual

These commands illustrate how to manage Git worktrees using OB1's built-in functionality. You can create new worktrees for specific branches and list all currently active worktrees.

```bash
/worktree new auth-refactor
```

```bash
/worktree list
```

--------------------------------

### Configure OB1 Security Settings

Source: https://www.openblocklabs.com/manual

Enables the enforceWorktree setting to prevent direct file modifications on the main branch, requiring a worktree for edits.

```json
// ~/.ob1/settings.json
{
  "security": {
    "enforceWorktree": true
  }
}
```

--------------------------------

### JSON Configuration for BeforeTool Hook

Source: https://www.openblocklabs.com/manual

This JSON configuration sets up a 'BeforeTool' hook that executes a shell script located at '.ob1/hooks/block-secrets.sh' before 'write_file' or 'replace' operations.

```json
{
  "hooks": {
    "BeforeTool": [
      {
        "matcher": "write_file|replace",
        "hooks": [
          {
            "type": "command",
            "command": ".ob1/hooks/block-secrets.sh"
          }
        ]
      }
    ]
  }
}
```

--------------------------------

### Define Skill Frontmatter

Source: https://www.openblocklabs.com/manual

Configuration block for custom skills, allowing specification of metadata such as the preferred AI model for execution.

```yaml
---
name: code-review
model: claude-opus
description: In-depth code review with architectural analysis
---

# Code Review Skill
...
```

--------------------------------

### Search Threads in OB1

Source: https://www.openblocklabs.com/manual

These JSON snippets demonstrate how to search through past conversation threads in OB1. You can search within the current project, across all projects, or retrieve the full content of a specific thread by its ID.

```json
{
  "action": "list",
  "search": "auth refactor"
}
```

```json
{
  "action": "list",
  "scope": "all",
  "search": "database migration"
}
```

```json
{
  "action": "get",
  "thread_id": "3"
}
```

--------------------------------

### Manage Subagent Assignments with OB1 CLI

Source: https://www.openblocklabs.com/manual

Commands to view, set, and reset subagent assignments in OB1. These commands allow users to inspect current model assignments for subagents, override specific agents with different models for a session, and clear all custom overrides.

```bash
/subagents                          # View current assignments
/subagents set explore gpt-5        # Override explore agent
/subagents reset                    # Clear all overrides
```

--------------------------------

### Shell Script to Detect Secrets

Source: https://www.openblocklabs.com/manual

This shell script checks input content for common secret patterns like API keys or passwords using grep. It outputs a JSON indicating whether to deny or allow based on the detection.

```shell
if echo "$content" | grep -qE 'api[_-]?key|password|secret'; then
  echo '{"decision": "deny", "reason": "Potential secret detected"}' 
  exit 0
fi

echo '{"decision": "allow"}'
exit 0
```
Tokens:
Raw