Skill UI

This skill guides the implementation of continuous, risk-adaptive identity verification crucial for Zero Trust Architecture (ZTA). It covers deploying phishing-resistant MFA (FIDO2/WebAuthn), advanced risk-based conditional access policies, and identity governance strategies. Users learn how to move beyond traditional credentials by leveraging multiple signals—such as device posture, behavioral biometrics, and location context—to dynamically assess and enforce trust levels before granting access.

This guide details implementing Zero Trust Network Access (ZTNA) across major cloud providers (AWS, Azure, GCP). It replaces traditional VPNs by enforcing identity-based, context-aware access controls using identity-aware proxies and micro-segmentation. Use it to secure cloud workloads, limit lateral movement, and comply with modern security standards like BeyondCorp.

Provides a comprehensive guide to implementing the Zero Trust security model using Google BeyondCorp Enterprise. It configures the Identity-Aware Proxy (IAP) to enforce access decisions based on granular context (user identity, device posture, network attributes), moving away from traditional network perimeters. Ideal for securing sensitive GCP resources and meeting compliance requirements.

A persistent file-based planning system for multi-step AI agent work. Maintains task_plan.md, findings.md, and progress.md on disk with lifecycle hooks to inject context. Automatically resumes from saved plans. Suitable for research or tasks requiring more than 5 tool calls.

Persistent file-based planning for multi-step AI agent work. Keeps task_plan.md, findings.md and progress.md on disk so plan context survives long sessions, with lifecycle hooks injecting the active plan and optional local session catchup.

This is an advanced, structured workflow designed to conduct deep, multi-source investigations. It synthesizes current public evidence with user-provided context to generate highly structured reports. The workflow explicitly differentiates between verifiable facts, user-supplied evidence, logical inferences, and actionable recommendations, making it ideal for complex comparative analysis or continuous monitoring tasks.

Fetches Google Docs body text and uses it as the Gmail message content so you can quickly turn a document into an email draft for newsletters or updates.

Copilot Spaces allow users to consolidate and manage project-specific knowledge, including documentation, code snippets, architectural standards, and operational instructions. By loading a Space, Copilot grounds its responses not just on general knowledge, but on the curated, authoritative context of the entire project, enabling accurate Q&A, code generation, and workflow execution based on team standards. Ideal for onboarding, large-scale projects, and maintaining consistency.

Persistent file-based planning for multi-step AI agent tasks. Maintains task_plan.md, findings.md, and progress.md on disk, with lifecycle hooks injecting selected plan context. Best for research or work needing 5+ tool calls.

Converts text-based documentation into structured Markdown by following explicit instructions, documented options, or reference files, supporting header levels, patterns, platforms, and cleanup steps for polished technical documents.

A skill for continuous file-based planning in multi-step AI agent workflows. Maintains task_plan.md, findings.md, and progress.md on disk, with lifecycle hooks injecting project-specific planning context. Supports explicit session catch-up via session-catchup.py for metadata or limited replay, and optional controlled mode. Designed for research or tasks requiring 5+ tool calls.

A skill for persistent, file-based planning in multi-step AI agent work. It maintains task_plan.md, findings.md, and progress.md on disk, with lifecycle hooks to inject planning context. Automatic recovery reads only project planning files. Optional gate mode can request continuation only with host support and never executes commands from Markdown. Suitable for research or tasks with more than 5 tool calls.