birdwatch

Monitor live systems and verify API/DB state during testing sessions.

Updated Apr 26, 2026
One-click install
npx skills add https://github.com/kennykankush/skillpack --skill birdwatch
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: birdwatch
Source: https://github.com/kennykankush/skillpack/tree/main/plugins/agents/skills/birdwatch
Command: npx skills add https://github.com/kennykankush/skillpack --skill birdwatch

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Birdwatch enables agents to act as evidence-backed observers during live testing, turning noisy, uncertain sessions into structured insight by tracking running state, capturing logs, and surfacing discrepancies across surfaces.

Core Features & Use Cases

  • Live observation of apps, services, and infra during test sweeps to keep the user oriented with real-time evidence.
  • Verify UI against API/DB state by cross-checking outputs and highlighting mismatches.
  • Durable note-taking and experience dumps that can be drafted into issues or handoffs.
  • Host-agnostic operation across Codex, Claude Code, and other agent hosts with a clear evidentiary contract.

Quick Start

Ask your agent to start Birdwatch and begin a live observation session during a testing sweep.

Frequently Asked Questions about birdwatch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I capture evidence-backed logs during live application testing?

To capture evidence-backed logs during live application testing, you can use an observation skill that monitors running processes, verifies API and database state, and surfaces discrepancies in real time.

What is live system observation for QA and how does it work?

Live system observation for QA is the process of monitoring running apps and infrastructure during test sweeps to provide real-time evidence. It works by defining a fixed contract of watch scopes, allowed actions, and evidence sources to log findings.

Can I verify UI state against API and database state during a test sweep?

Yes, you can verify UI state against API and database state during a test sweep by cross-checking outputs and highlighting mismatches. The observation session tracks running processes and surfaces discrepancies across surfaces.

Does this live observation tool work across different agent hosts like Codex and Claude Code?

Yes, this live observation tool operates in a host-agnostic manner across Codex, Claude Code, and shell runners. It defines a fixed contract of inputs and outputs so it can function consistently across different agent environments.

How do I start a live observation session during infra and product testing?

To start a live observation session during infra and product testing, ask your agent to begin monitoring and define the watch scope, allowed actions, evidence sources, and output targets. This establishes the fixed contract for the testing sweep.

What is the best way to turn noisy testing session logs into structured issues and handoffs?

The best way to turn noisy testing session logs into structured issues is to use durable note-taking and experience dumps during live observation. This captures evidence-backed findings that can be drafted directly into issues or handoffs.