
from agent-messenger69
Zero-config CLI that lets AI agents and humans read and send Channel Talk messages, list chats/groups, and manage multi-workspace access by extracting desktop/b
Agent ChannelTalk provides a TypeScript CLI that enables AI agents (and humans) to interact with Channel Talk workspaces. It auto-extracts credentials from the Channel Talk desktop app with a Chromium browser fallback, and exposes commands to snapshot workspaces, list chats and groups, send messages, and search messages. Designed for automation and integration workflows where agents need read/write access to support conversations.
Use this skill when an agent must act as a manager in Channel Talk: generate workspace snapshots, post messages to groups or user chats, search historical messages, or programmatically inspect group/user-chat lists. It is useful for automation, monitoring, ticket summarization, and programmatic responses where zero-config authentication is desired.
templates/ (examples: post-message.sh, monitor-chat.sh, workspace-summary.sh) — has_references=true so implementation notes exist.--full vs default), and output formats (JSON/pretty).Primarily built for shell/Bash-driven integrations and AI agents that can call binaries (Copilot/Codex-like tooling). Works where agents can run CLI tools or call the installed agent-channeltalk binary.
Agent ChannelTalk is a zero-config CLI for interacting with Channel Talk workspaces, extracting credentials from the desktop app or browser automatically. The SKILL.md is thorough with clear command references, auth workflows, memory management guidance, and a helpful feature comparison table. No bundled scripts to test — it's primarily a CLI wrapper skill. Well-documented for its niche.
No security concerns. Credential extraction is the core design and is handled responsibly (0600 file permissions, no hardcoded secrets, explicit guidance against storing sensitive data in memory). Niche but well-executed skill for Channel Talk users.