obscuritylabs/colossus

Browse and interact with websites through Colossus browser tools, including local previews, forms, screenshots, and page inspection. Use when the task requires an interactive browser.

Implement features, fix bugs, debug failures, or refactor software in an existing repository or a new project. Use when a coding request needs a working change and verification of the requested behavior.

Read and operate Windows desktop apps through Colossus computer tools, using accessible controls, window screenshots, mouse, and keyboard. Use when the task requires a native app rather than a website or repository tool.

Write or edit developer documentation using selectively loaded guidance based on the Google Developer Documentation Style Guide. Use for READMEs, API documentation, tutorials, and how-to guides.

Create or update portable Agent Plugins with Agent Skills or MCP servers, validate them with Colossus, and package them as OCI artifacts for registry or offline distribution.

Create, inspect, pause, enable, or delete recurring plain-language tasks and workflow schedules with exact local timing, workspace permissions, and durable confirmation.

Perform evidence-based security reviews of repositories, changes, or components. Map threats and trust boundaries, trace attacker-controlled inputs, verify findings, and use bounded subagents and durable working notes for large reviews. Use for explicit security review, vulnerability analysis, or security audit requests, not ordinary code review.