Gemini CLI

Chat

Google

Product overview

Name of Agent: Gemini CLI
Short description of agent: "Query and edit large codebases, generate apps from images or PDFs, and automate complex workflows—all from your terminal with Gemini 3" (link, archived)
Date of release: 25/06/2025 (link, archived)
Advertised use: "complete complex use cases like fixing bugs, creating new features, and improving test coverage", but also "versatile local utility that you can use for a wide range of tasks, from content generation and problem solving to deep research and task management" (source, archived)
Monetisation/Usage price: Free, 20, higher rate limits (link, archived) 125,
Who is using it?: end user and enterprise customers for coding and prototyping
Category: Chat

Company & accountability

Developer: Google
Name of legal entity: Google LLC (link, archived)
Place of legal incorporation: Delaware, USA (link, archived)
For profit company?: Yes
Parent company?: Alphabet Inc
Governance documents analysis: Google TOS (link, archived), GenAI Prohibited Use Policy (link, archived), Gemini CLI (link, archived)
AI safety/trust framework: Gemini App safety and policy guidelines (link, archived), Responsible AI Principles (link, archived)
Compliance with existing standards: Looks like mentions of Gemini and compliance standards are for enterprise versions of them, but unclear (link, archived)

Technical capabilities & system architecture

Model specifications: Powered by Gemini models. Default, "gemini-3-pro-preview (if enabled), gemini-3-flash-preview (if enabled)". You can select which Gemini model to use. (source, archived)
Observation space: File System Tools: For interacting with files and directories (reading, writing, listing, searching, etc.). Shell Tool (run_shell_command): For executing shell commands. Web Fetch Tool (web_fetch): For retrieving content from URLs. Web Search Tool (google_web_search): For searching the web. Memory Tool (save_memory): For saving and recalling information across sessions. To do Tool (write_to dos): For managing subtasks of complex requests. MCP servers: MCP servers act as a bridge between the Gemini model and your local environment or other services like APIs. (source, archived)
Action space: File System Tools: For interacting with files and directories (reading, writing, listing, searching, etc.). Shell Tool (run_shell_command): For executing shell commands. MCP servers: MCP servers act as a bridge between the Gemini model and your local environment or other services like APIs. (source, archived)
Memory architecture: memory file is a plain text Markdown file, so you can view and edit it manually if needed. It is used through the memory tool (source , archived)more details on how it is setup (link, archived)
User interface and interaction design: Chabot in CLI
User roles: Operator (issues queries, which the agent to responds to); Executor (user may take actions/make decisions based on outputs); Examiner (user can use thumbs up/down buttons to give feedback)
Component accessibility: Open source (link), for the agentic harness, the models are closed source

Autonomy & control

Autonomy level and planning depth: L1-L3: Tasks that users assigns to the agent are often narrow in scope. More complex tasks need multi-turn conversations where the user is in charge of planning. Agent always comes back to the user and awaits further instructions. Agent can create a plan and execute on it across multiple steps.
User approval requirements for different decision types: Yes, permission for running bash commands or editing files (source, archived)
Execution monitoring, traces, and transparency: Visible (albeit summarized) CoT with a description of current task being executed
Emergency stop and shut down mechanisms and user control: User can pause/stop the agent at any time
Usage monitoring and statistics and patterns: User can see how much context is used

Ecosystem interaction

Identify to humans?: None found
Identifies technically?: None found, but presumably the same as Gemini
Interoperability standards and integrations: MCP (link, archived)
Web conduct: None found, Gemini CLI documentation for web_fetch does not mention checking or complying with robots.txt for direct fetches (robots.txt compliance is not specified) (link, archived)

Safety, evaluation & impact

Technical guardrails and safety measures: Default restriction in what tools can do without approval. When a folder is untrusted, the Gemini CLI runs in a restricted “safe mode” to protect you. In this mode, the following features are disabled list (link, archived)
Sandboxing and containment approaches: All tools are subject to restrictions enforced by sandboxing, (source, archived)
What types of risks were evaluated?: Underlying models are evaluated in agentic settings, see model cards (link, archived), e.g. computer use (link, archived)
(Internal) safety evaluations and results: Underlying models are evaluated in agentic settings, see model cards (link, archived), e.g. computer use (link, archived)
Third-party testing, audits, and red-teaming: None found
Benchmark performance and demonstrated capabilities: Refer to model cards (link, archived), e.g. computer use (link, archived)
Bug bounty programmes and vulnerability disclosure: Yes (link, archived)
Any known incidents?: None found
← Back to 2025 Index