references/get-tasking.md
GitHubGet Tasking
This page describes the protocol for an agent to request pending tasks from Mythic.
Request: Agent -> Mythic
Base64( CallbackUUID + EncBlob(
JSON({
"action": "get_tasking",
"tasking_size": 1,
// optional - forward P2P delegate messages
"delegates": [
{"message": "<agentMessage>", "c2_profile": "ProfileName", "uuid": "uuid here"}
],
// optional - defaults to true
"get_delegate_tasks": true
})
)
)
Fields
| Field | Required | Default | Description |
|---|---|---|---|
action |
Yes | - | Must be "get_tasking" |
tasking_size |
No | 1 | Max number of tasks to return. Use -1 for ALL pending tasks. |
delegates |
No | - | Array of P2P forwarded messages from linked agents |
get_delegate_tasks |
No | true |
If true, also fetch tasks for agents reachable through this callback |
About get_delegate_tasks
When set to true (default), if agentA has a route to agentB and agentB has a pending task, agentA's get_tasking will also pick up agentB's task. Set to false if your linked agents issue their own periodic get_tasking messages - prevents the parent from consuming a child's tasks.
Response: Mythic -> Agent
Base64( CallbackUUID + EncBlob(
JSON({
"action": "get_tasking",
"tasks": [
{
"command": "command name",
"parameters": "command param string",
"timestamp": 1578706611.324671,
"id": "task uuid"
}
],
// P2P responses for delegates in the request
"delegates": [
{"message": "<agentMessage>", "c2_profile": "ProfileName", "uuid": "uuid here"}
]
})
)
)
Fields
| Field | Description |
|---|---|
tasks |
Array of 0 to tasking_size task objects |
tasks[].command |
The command name to execute |
tasks[].parameters |
String containing command parameters. If the command has structured parameters like {"remote_path": "/tmp/file", "file_id": "uuid"}, this will be a JSON string that the agent must parse. |
tasks[].timestamp |
Unix timestamp for ordering |
tasks[].id |
Task UUID - used when posting responses |
delegates |
Responses for any delegate messages sent in the request |
Combining with Responses
The get_tasking request CAN also include these fields:
responses- task output (same format aspost_response)socks- SOCKS5 proxy data. Array of{exit, server_id, data}messages for tunneling TCP connections through the agent. See SOCKS.rpfwd- Reverse port forward data. Array of{exit, server_id, data, port}messages for reverse-tunneling connections back through Mythic. See Reverse Port Forward.edges- P2P connection updatesalerts- Alert messagesinteractive- Interactive tasking data
Important: Mythic may include socks and/or rpfwd arrays in its response to ANY get_tasking request. The agent must always check for and process these arrays, not just when it expects proxy traffic.
This allows the agent to send results AND request new tasking in a single message, avoiding two round-trips per sleep cycle.
Example combined message:
{
"action": "get_tasking",
"tasking_size": -1,
"responses": [
{
"task_id": "uuid-of-completed-task",
"completed": true,
"user_output": "command output here"
}
]
}