Skip to content

Pepebot v0.5.2

Choose a tag to compare

View all tags
@github-actions github-actions released this 22 Feb 06:34
· 122 commits to main since this release
v0.5.2
064ed1a
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

🐸 Pepebot v0.5.2 - Workflow CLI

Release Date: 2026-02-22

🎉 What's New

📋 Workflow CLI — Run Automations Without the Agent

You can now manage and execute workflows directly from the terminal! No need to chat with the agent just to run an automation.

# List all your workflows
pepebot workflow list

# Inspect a workflow before running it
pepebot workflow show my_workflow

# Run a workflow from your workspace
pepebot workflow run my_workflow

# Run with variable overrides
pepebot workflow run my_workflow --var device=emulator-5554 --var query=hello

# Run directly from a JSON file (great for testing new workflows)
pepebot workflow run -f /tmp/test_workflow.json
pepebot workflow run -f /tmp/test_workflow.json --var key=value

# Validate a workflow's structure
pepebot workflow validate my_workflow
pepebot workflow validate -f /tmp/test_workflow.json

# Delete a workflow
pepebot workflow delete old_workflow

The -f flag lets you point at any JSON file on disk, bypassing the workspace — perfect for iterating on new workflows before saving them to your workspace.

Note: Workflows with goal or agent steps that require an LLM will return a note explaining they need the full agent loop. Use pepebot agent -m "run workflow X" for those.

🐸 Pepebot v0.5.1 - Workflow Skill & Agent Steps + ADB Activity Recorder

Release Date: 2026-02-21

🎉 What's New

🧩 Workflow Skill Steps

Workflows can now load skill content and combine it with goal instructions! Use the skill field to bring specialized knowledge into your workflow steps.

{
  "name": "analyze_with_skill",
  "skill": "workflow",
  "goal": "Using this skill's knowledge, analyze the data from {{collect_output}}"
}

🤝 Workflow Agent Steps

Delegate workflow goals to other registered agents! Use the agent field to leverage different models and prompt configurations within a single workflow.

{
  "name": "research",
  "agent": "researcher",
  "goal": "Research {{topic}} and provide a summary"
}

Agent responses are stored as {{step_name_output}} for use in subsequent steps.

🔄 Graceful Gateway Restart

Restart the gateway without killing the process! Config changes, new agents, updated API keys — all take effect with a simple restart.

Via chat command (any channel):

/restart

Via API:

curl -X POST http://localhost:18790/v1/restart

Via signal:

kill -HUP $(pgrep pepebot)

All services (HTTP server, channels, cron, heartbeat) are gracefully stopped and re-initialized from fresh config.

🎬 ADB Activity Recorder

Record your Android device interactions and automatically generate replayable workflow files! Simply use your device while Pepebot watches — taps and swipes are captured in real-time via ADB.

How to use:

User: "Record my Android actions as a workflow named login_flow"
  • Pepebot starts listening to touch events on your device
  • Perform taps, swipes, and other interactions normally
  • Press Volume Down to stop recording
  • A workflow JSON file is automatically generated and saved

What gets captured:

  • Tap gestures (short touch, small movement)
  • Swipe gestures (long touch with displacement)
  • Final screenshot and UI dump for verification

Generated workflows use standard adb_tap and adb_swipe steps with a {{device}} variable, so you can replay on any connected device.

Key Features

  • Real-time touch event streaming via getevent -l
  • Automatic input device discovery and screen resolution detection
  • Smart gesture classification with debounce filtering
  • Post-recording screenshot + UI dump for LLM-based verification
  • Standard workflow format — edit, extend, or chain with other workflows

📦 Installation

Using Install Script (Recommended)

curl -fsSL https://raw.githubusercontent.com/pepebot-space/pepebot/main/install.sh | bash

Using Homebrew

brew tap pepebot-space/pepebot
brew install pepebot

Using Docker

docker pull ghcr.io/pepebot-space/pepebot:latest
docker run -it --rm pepebot:latest

Manual Download

Download the appropriate binary for your platform from the releases page.

🚀 Quick Start

  1. Connect your Android device via USB (enable USB debugging)
  2. Start recording: 
    pepebot agent -m "Record my actions as a workflow named my_recording"
  3. Interact with your device (tap, swipe)
  4. Press Volume Down to stop
  5. Replay: 
    pepebot agent -m "Execute workflow my_recording"

📚 Documentation

🔗 Links

📝 Full Changelog

For a complete list of changes, see CHANGELOG.md.

Note: When upgrading from v0.5.0, all existing configurations, workflows, and data are preserved. No migration needed.

📦 Platform-Specific Downloads

Choose the appropriate binary for your system:

Linux

  • x86_64: pepebot-linux-amd64.tar.gz
  • ARM64: pepebot-linux-arm64.tar.gz
  • ARMv7: pepebot-linux-armv7.tar.gz (Raspberry Pi 3/4)
  • ARMv6: pepebot-linux-armv6.tar.gz (Raspberry Pi 1/Zero)
  • RISC-V 64: pepebot-linux-riscv64.tar.gz
  • MIPS: pepebot-linux-mips.tar.gz / pepebot-linux-mipsle.tar.gz
  • MIPS64: pepebot-linux-mips64.tar.gz / pepebot-linux-mips64le.tar.gz

macOS

  • Intel: pepebot-darwin-amd64.tar.gz
  • Apple Silicon (M1/M2/M3): pepebot-darwin-arm64.tar.gz

Windows

  • x86_64: pepebot-windows-amd64.tar.gz
  • ARM64: pepebot-windows-arm64.tar.gz

BSD

  • FreeBSD x86_64: pepebot-freebsd-amd64.tar.gz
  • FreeBSD ARM64: pepebot-freebsd-arm64.tar.gz

Android

  • ARM64: pepebot-android-arm64.tar.gz (Most devices, Android 5.0+)

🔐 Verification

All binaries include SHA256 checksums. Verify your download:

sha256sum -c pepebot-*.tar.gz.sha256

Built with ❤️ using Go

Assets 34
Loading