Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.
Use when asked to implement, build, fix, or create features - especially before writing any code or when scope and requirements are unclear
Use when orchestrating complex multi-step tasks requiring delegation, parallelization, or systematic completion verification - especially when tempted to do everything yourself or ask user codebase questions
Use when facing trade-offs, subjective judgments, uncertain decisions, or when diverse viewpoints would improve judgment quality. Triggers include "council", "다른 의견", "perspectives", "what do others think".
Code review orchestration skill - fans out angle finders across review angles and merges their raw candidate findings
Autonomous objective-pursuit executor — decomposes an objective into Six Slots and a Story set, dispatches execution to sisyphus, then re-pursues the objective across plan/execute cycles until an objective-level completion gate confirms the verification surface is met.
Autonomous multi-story objective-pursuit executor — restructures goal into a per-story sequential loop. Auto-generates and bulk-approves a Story set, dispatches each confirmed story to sisyphus one at a time gated by a per-story APPROVE verdict, then runs one independent code-review over the accumulated diff at the final story.
Socratic deep interview with mathematical ambiguity gating before autonomous execution