Gemini CLI 0.26 Adds Hooks and Skills for Agent Customization

Srinath Padmanabhan

Jan 28, 2026 · Updated Apr 25, 2026

Gemini CLI v0.26 adds Hooks and Agent Skills. Skills extend the agent with domain expertise that loads on demand, while Hooks give you middleware-style control over every agent action - validating outputs, injecting context, or enforcing policies.

Gemini CLI, Google's open-source agentic coding CLI, ships Hooks and Agent Skills in v0.26. Hooks are scripts that execute at lifecycle events like BeforeTool or AfterAgent. They work as middleware: intercept any agent action to validate outputs, inject data, enforce policies, or trigger automations. Full control over the agentic loop without modifying source code.

Agent Skills are markdown files in .gemini/skills/ that package domain expertise, procedural workflows, and task-specific resources. The agent activates them only when your prompt needs that knowledge, keeping context focused instead of loading everything upfront.

The release also adds /rewind for rolling back to previous checkpoints when something goes wrong, plus persistent permission policies. Update with npm i -g @google/gemini-cli@latest.

View the full update on developers.googleblog.com

Srinath Padmanabhan

@SriThreePOJan 28

Here is Gemini CLI’s January 26th weekly update for v0.26.0 - 🧩Skills: Introducing Agent Skills that allow you to extend Gemini CLI with specialized expertise, procedural workflows, and task-specific resources. (vid, pr by NTaylorMullen) - Introduced the built-in skill-creator for generating tools - Added security consent prompts to ensure safer, easier management of agent capabilities. - 🪝Hooks: Now officially enabled by default, hooks provide a way to fully control and customize the agentic loop. (gif, work by Edilmo and abhipatel12) https://t.co/zF0eqckTsr - 🎉Gemini CLI Extensions: - Supabase https://t.co/NGGwuEQCd2 - Connect AI assistants directly to Supabase projects to manage tables, run migrations, and query data seamlessly without context switching. - ⏪New /rewind Command: Implemented the /rewind slash command, including a confirmation dialog and a new viewer component for navigating past states. If something goes wrong you now have the ability to revert both history and changes to a previous point in time. - Persistent "Always Allow" Policies: Users can now save "Always Allow" decisions for tool executions, with granular control over specific shell commands and multi-cloud platform tools. - Folder Trust: - Opening Gemini CLI in a folder for the first time will now surface a trust dialog - Enable or disable checking folder trust holistically with folderTrust in your settings.json. Discussion Post: https://t.co/cdnWW3lyVW 🧵

19147

View on X

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Keep reading

Google Launches Gemini API Skills to Fix Outdated AI Coding Knowledge

Google introduced Gemini API Skills, a lightweight framework that provides AI agents with real-time SDK knowledge and documentation. By using the new Gemini API developer skill, the Gemini 3.1 Pro Preview model increased its coding task success rate from 28.2% to 96.6%. This shift demonstrates that providing agents with structured, up-to-date primitives is more effective than relying on static model weights for rapidly evolving software tools.

Google Labs Releases Agent Skills Library for Stitch MCP Server

Tom DörrMar 25

Google Labs Releases Agent Skills Library for Stitch MCP Server

Google Labs open-sourced stitch-skills, a library of agent skills for the Stitch MCP server following the Agent Skills open standard. Each skill installs with a single CLI command and runs across coding agents including Gemini CLI, Claude Code, and Cursor.

Google GemmaMay 1

Google Gemini CLI Integrates Local Gemma Models for Intelligent Task Routing

Gemini CLI v0.40.0 introduces experimental support for running Gemma models locally to handle intelligent routing decisions. By offloading intent analysis to the user's hardware, the agent reduces cloud API dependency and latency for simple tasks. This marks the first step toward a roadmap of full local execution for Google's terminal-based agent.

OpenClaw 2026.3.11 Adds Free 1M Context Models and Gemini Memory

OpenClawMar 18

OpenClaw 2026.3.11 Adds Free 1M Context Models and Gemini Memory

OpenClaw 2026.3.11 ships Hunter Alpha and Healer Alpha — two free stealth models with 1M context via OpenRouter — available for roughly one week. The release also adds gemini-embedding-2-preview for memory search and a new OpenCode Go provider.