gemini-runtime
Controls Gemini CLI sessions for automated model interaction.
Install
mkdir -p .claude/skills/gemini-runtime && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/13638" && unzip -o skill.zip -d .claude/skills/gemini-runtime && rm skill.zipInstalls to .claude/skills/gemini-runtime
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Gemini CLI runtime/session control: use Gemini CLI slash commands and CLI flags to manage sessions, models, and restarts in a worker terminal. Trigger when the controller needs to operate the Gemini CLI like a human.Key capabilities
- →Manage Gemini CLI session state
- →Select Gemini models
- →Reset current conversations
- →Rewind to previous session steps
- →Resume saved sessions
- →List and delete sessions
How it works
This skill operates the Gemini CLI by executing slash commands and CLI flags to manage session state, model selection, and restarts.
Inputs & outputs
When to use gemini-runtime
- →Resetting chat sessions
- →Switching Gemini models
- →Resuming previous sessions
About this skill
Gemini Runtime
Overview
Operate the Gemini CLI safely: manage session state, model selection, and restarts from a worker terminal.
Session Safety
- Confirm idle state
- Snapshot and/or
statusthe worker; do not intervene mid-run. - Only proceed when the worker is at a prompt or explicitly idle.
- Prefer in-session resets
- Use
/clearto reset the current conversation. - Use
/rewindto go back to a previous step. - Use
/statsor/contextto confirm state.
- Full reset (only between tasks)
- Use
/quit(or/exit) to leave the CLI. - Relaunch with appropriate flags when a fresh context is needed.
Core Slash Commands (Gemini CLI)
Use /help to see the full list. Common commands:
/modelselect a model./resumeresume a saved session./restorerestore a previous state./chatmanage chat sessions./clearreset the current conversation./contextshow current context./statsshow session stats./settingsopen settings./privacyview privacy info./policiesview policies./bugfile a bug./vimtoggle vim mode./rewindstep backward in the session./quitor/exitexit the CLI.
Resume / Continue (Shell)
gemini --resume <SESSION_ID>orgemini -r <SESSION_ID>resume a session.gemini --list-sessionslist available sessions.gemini --delete-session <SESSION_ID>delete a session by index.
Guardrails
- Do not restart mid-run.
- Prefer
/clearor/rewindover full restarts when possible. - Use
/statsor/contextafter model changes to confirm state. - If the worker is not Gemini CLI, switch to the model-specific runtime skill instead.
When not to use it
- →When the worker is not Gemini CLI
- →When intervening mid-run of the worker
- →When a full reset is needed between tasks
Limitations
- →Requires confirmation of idle state before intervention
- →Prefers in-session resets over full restarts
- →Must use `/stats` or `/context` to confirm state after model changes
How it compares
This workflow provides a controlled and safe way to interact with the Gemini CLI, ensuring session integrity and preventing unintended interruptions, unlike direct, unmanaged CLI usage.
Compared to similar skills
gemini-runtime side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| gemini-runtime (this skill) | 0 | 5mo | No flags | Intermediate |
| autonomous-loops | 0 | 2mo | Review | Advanced |
| continuous-learning | 0 | 1mo | Review | Advanced |
| mcporter | 0 | 6mo | Review | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
autonomous-loops
Rajbharti06
Self-improvement loop patterns — eval → learn → improve cycles running without human input
continuous-learning
mostafa-ismail-2004
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents. Features project-scoped instincts to prevent cross-project contamination.
mcporter
consuelohq
--- name: mcporter description: Call MCP servers via MCPorter - browser automation, vision, and other MCP tools homepage: https://github.com/steipete/mcporter metadata: { "openclaw": { "emoji": "🧳", "requires": { "bins": ["npx"] } } }
openai-whisper
qianleigood
Use when transcribing local audio or video on Apple Silicon without an API key, especially when MLX Whisper should be the default general-purpose local speech-to-text path.
opencode-cli
SpillwaveSolutions
This skill should be used when configuring or using the OpenCode CLI for headless LLM automation. Use when the user asks to "configure opencode", "use opencode cli", "set up opencode", "opencode run command", "opencode model selection", "opencode providers", "opencode vertex ai", "opencode mcp servers", "opencode ollama", "opencode local models", "opencode deepseek", "opencode kimi", "opencode mistral", "fallback cli tool", or "headless llm cli". Covers command syntax, provider configuration, Vertex AI setup, MCP servers, local models, cloud providers, and subprocess integration patterns.
robotics-code-generator
HumaizaNaz
Generates clean, runnable ROS 2, Gazebo, Isaac Sim, and VLA code for humanoid robotics