Podcvd

podcvd Skill Instructions

This skill fully orchestrates the lifecycle of Cuttlefish instances and their groups within container environments, including the automatic establishment and teardown of ADB connections.

> [!WARNING]

> Exclusive ADB Connection Control: This skill internally manages all required ADB connections when creating, removing, or clearing instances. You MUST NOT invoke external skills, MCP servers, or secondary tools to establish ADB connections to these Cuttlefish instances, as doing so will cause system conflicts.

> [!WARNING]

> Strict Isolation from cvd: podcvd and standard cvd do not share state or information. Mixing their usage will cause inconsistent states and system errors.

> * MUST NOT use standard cvd commands if this podcvd skill has been used in the environment.

> * MUST NOT use this podcvd skill if standard cvd was previously used to manage instances.

> * Why Instances May Seem Missing: Because they do not share state, instances created via standard cvd will not be visible to podcvd (e.g., podcvd fleet will show no instances), and vice versa. If you cannot find an expected Cuttlefish instance, verify which tool was used to create it.

Prerequisites

• Required Build Artifacts: Launching Cuttlefish instances requires appropriate Cuttlefish host tools (host binaries and libraries) and Android device images (such as system.img, boot.img, and vendor.img):
• Inside an Android Workspace: You must initialize the environment (e.g., lunch <target>) and build the required targets (e.g., m) to ensure that valid Cuttlefish host tools (inside ANDROID_HOST_OUT) and Android device images (inside ANDROID_PRODUCT_OUT) are generated.
• Outside an Android Workspace: If you are not in an active build environment, you must ensure that compatible, pre-built Cuttlefish host tools and Android device images are already present in the working directory or are otherwise accessible via the relevant environment paths (like ANDROID_HOST_OUT and ANDROID_PRODUCT_OUT).

Setup & Installation Guidance

• Missing podcvd Command: If podcvd is not installed (e.g., which podcvd fails), install the package using sudo apt install cuttlefish-podcvd -y.
• Host Initialization Required: If executing podcvd fails due to an unconfigured host (e.g., explicit prompts to run podcvd-setup or missing device permissions such as /dev/kvm), execute podcvd-setup to initialize the host.

Mandatory Environment Configuration (Client ID Tracking)

Setting the PODCVD_CLIENT_ID environment variable is MANDATORY whenever executing podcvd subcommands. This ensures proper session tracking and prevents resource collisions.

Execution Rule (Mandatory)

• You MUST check if the reference file references/podcvd_client_id.md exists within this skill directory.
• If references/podcvd_client_id.md exists: You MUST read its contents using view_file and strictly follow the platform-specific instructions provided inside to construct and pass PODCVD_CLIENT_ID before calling podcvd.
• If references/podcvd_client_id.md does NOT exist: Only as an exception, omit setting PODCVD_CLIENT_ID (leave it empty) and execute podcvd directly.

Key Execution Guidelines & Considerations (Crucial)

> [!IMPORTANT]

> Always follow these essential execution policies when managing Cuttlefish instances:

• ADB Connection Lifecycle: Subcommands (e.g., create, start, restart, powerwash) automatically establish or re-establish ADB connections, while subcommands (e.g., stop, remove, clear, reset) automatically tear down ADB connections. Never invoke external ADB tools manually.
• Dynamic Group Identifier Tracking: Executing podcvd create initializes a new instance group and returns a JSON response containing its assigned "group_name". Because "group_name" is the primary identifier for managing that group, you MUST extract and track this string (from the create output or podcvd fleet) to dynamically pass --group_name=<group_name> in subsequent lifecycle operations (e.g., stop, start, restart, remove).
• Booting & Execution Duration: Starting or booting Cuttlefish instances (e.g., executing create, start, restart, powerwash) takes approximately 1 minute for the virtual device boot process and ADB connection setup to complete. During this boot window, podcvd fleet will report status "Starting" and adb devices -l may temporarily show an empty list. You MUST wait for podcvd fleet status to transition to "Running" (up to 1-2 minutes) before verifying ADB device readiness or executing teardown operations. Do not prematurely cancel, clear, or report failure while status is still "Starting".
• Status Verification: Always verify instance state using podcvd fleet BEFORE and AFTER state transitions. Before issuing any state-changing subcommand (e.g., stop, restart, remove), execute podcvd fleet to ensure previous operations have finished and the group is in the expected state (e.g., "Running"). After transitions, verify status and confirm ADB connection readiness using podcvd fleet and adb devices -l.

Standard Ready-to-Use Commands (Recommended)

To ensure high success rates and avoid malformed options, prefer using the fully constructed standard command lines below for general tasks. Always include PODCVD_CLIENT_ID="<client_id>" at the start of every command line.

1. Create Instance Group (Establishes ADB Connection)

Creates an instance group with standard configurations and automatically establishes ADB connections. (For bulk creation, see the concurrency policy below).

PODCVD_CLIENT_ID="<client_id>" podcvd create --vhost_user_vsock=true --report_anonymous_usage_stats=n

2. List Active Fleets

Lists all currently running Cuttlefish instance groups.

PODCVD_CLIENT_ID="<client_id>" podcvd fleet

3. Remove Specific Group (Destroys ADB Connection)

Removes a targeted instance group and automatically destroys corresponding ADB connections. Replace <group_name> with the actual ID.

PODCVD_CLIENT_ID="<client_id>" podcvd --group_name=<group_name> remove

4. Clear All Groups (Destroys All ADB Connections)

Completely tears down all Cuttlefish instance groups and destroys all relevant ADB connections globally.

PODCVD_CLIENT_ID="<client_id>" podcvd clear

Concurrency & Parallel Execution Policy (Crucial)

Unless the user explicitly requests sequential execution, operations targeting different instance groups (such as creating multiple instances or removing multiple distinct groups) MUST ALWAYS be executed in parallel.

User intent is paramount and overrides this rule, but parallel processing is the mandatory default for any bulk workload.

> [!WARNING]

> Default Strict Prohibition: When handling bulk requests (e.g., "create 10 instances" or "remove groups A, B, and C"), you MUST NOT execute them sequentially by default. Waiting for one task to complete before initiating the next severely degrades system performance.

Execution Guidance for Agents:

Launch independent podcvd commands concurrently as background tasks, with leveraging agent's native capability dealing with concurrent behavior if possible.

Exceptions to Parallel Execution:

• User Override: The user explicitly asks to perform tasks sequentially (one by one).
• State Dependency: Operations sequentially target the exact same group_name (e.g., creating a group and immediately modifying it).
• Global Operations: The clear subcommand internally processes global cleanup across all groups, so manual parallelization does not apply.

Advanced Usage

Below are the currently supported podcvd subcommands in alphabetical order:

• bugreport: Captures diagnostic host logs into a zip archive for troubleshooting (optionally includes device adb bugreport)
• cache: Manage the files cached by cvd
• clear: Clears the instance database, stopping any running instances first
• create: Create a Cuttlefish instance group
• display: Enables hotplug/unplug of displays from running cuttlefish virtual devices
• env: Enumerate + Query APIs for all gRPC services made available by this virtual device instance
• fetch: Retrieve build artifacts based on branch and target names
• fleet: Lists active devices with relevant information
• help: Used to display help information for other commands
• lint: Error checks the input virtual device json config file
• load: Creates and starts an instance group from a JSON configuration file
• login: Acquire credentials
• logs: List and display log files
• monitor: Monitor device logs (launcher, kernel, and logcat) in real-time
• powerbtn: Trigger power button event on the device
• powerwash: Reset device to first boot state
• remove: Remove instance groups and related artifacts from the system
• reset: Remove all instance groups and kills orphaned cuttlefish processes
• restart: Restart device
• resume: Resume the cuttlefish device
• screen_recording: Record screen contents
• setup: Configure the host for Cuttlefish
• snapshot_take: Take snapshot of the cuttlefish device
• start: Start all Cuttlefish Instances in a group
• status: Query status of a single instance group. Use podcvd fleet for all devices
• stop: Stop Cuttlefish instances
• suspend: Suspend the cuttlefish device
• version: Prints version of cvd client and cvd server

To inspect supported flags and advanced setup options for any subcommand, query built-in help directly:

PODCVD_CLIENT_ID="<client_id>" podcvd help <subcommand>

Connection Details Formatting Tip (Mandatory Output Formatting)

Executing the create or fleet subcommands returns JSON outputs containing instance configuration details. When reporting these connection details in your responses, you MUST follow these exact structural formats:

• ADB Connection Address: You MUST combine the host IP address and the allocated ADB port into the standard IP:PORT format (for example: 192.168.112.1:6520). Never output only the port number or adb_port: 6520.
• Web UI Endpoint: You MUST strip away specific device paths from the web_access URL and present ONLY the base Protocol, IP, and Port (for example, if JSON returns "https://192.168.112.1:11443/devices/cvd_1-1-1/files/client.html", present it as https://192.168.112.1:11443).