BlueBrik1

Shadowqa — Communication skill for Claude Code

Communication community

Tool that syncs with Slack and Github for teams, or Claude and ChatGPT for individuals.

How to install Shadowqa

This entry records only its repository, not the path inside it, so there is no exact command to give. Open BlueBrik1/shadowqa and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Shadowqa does

Tool that syncs with Slack and Github for teams, or Claude and ChatGPT for individuals. Whether you use Slack, Claude Code, ChatGPT Chat, Codex, or Claude Chat, ShadowQA unifies context and generates dev plans, and hands them to your agent.

Alternatives in Communication

Prompts.chat — f.k.a. Awesome ChatGPT Prompts 154.3k ★
Lark Coding Agent Bridge — Bot that bridges Feishu/Lark messenger with a local Claude Code or Codex CLI 2.4k ★
Agentrove — A self-hosted web and macOS desktop app for Claude Code and Codex, powered by ACP 270 ★

README

◈ ShadowQA - https://youtu.be/M6zUsrPB5HA

**Team edition — Slack and GitHub context → Gemini plan → approval → isolated OpenCode session → independent checks → repair PR.**

**Individual edition — ChatGPT, Claude, Claude Code and Codex context → Gemini plan → approval → a session in the coding tool you already use → verified diff.**

**ShadowQA Live — the running web app → captured incident → Claude diagnosis (GPT fallback) → risk-graded patch → replay-verified fix → the same findings list, under the same automation mode.**

ShadowQA is a TypeScript service, a desktop app, a local Docker runner, a thin VS Code extension, and — for the individual edition — a Manifest V3 browser side panel and a local companion. Ordinary activity accumulates as source-linked context; pressing **Generate plan** in the desktop app starts planning. Gemini is the sole inference provider for planning in both editions.

This was originally built CLI-first (see the historical **REVISIONS TO PLAN** in [doc.md](doc.md)); the desktop app in [`desktop/`](desktop/) later replaced the terminal entirely as the product's interface — every command the CLI offered has a screen, and no manual `.env` editing or file-based configuration is part of the normal path anymore. §18 of [doc.md](doc.md) records that change and why.

Plans and checks cover the code. **ShadowQA Live** ([`live/`](live/README.md)) covers the running application: a browser SDK or Chrome extension watches clicks, requests, console output and exceptions on localhost, joins them into one incident, and a Python bridge diagnoses, patches, validates and replays the exact interaction before anything is called fixed. Live reports every incident to the ShadowQA service as a `detector: "live"` finding, obeys the project's automation mode, and takes approve/undo/pr/dismiss from the desktop app's Live tab, which also starts and stops the bridge process. It is the product's autonomous fixer for runtime bugs.

Both editions share one codebase,