references/file-uploads.md

GitHub

File Uploads (Mythic -> Agent)

This page describes how an agent pulls files from the Mythic server to the target.

Source: https://docs.mythic-c2.net/customizing/hooking-features/action-upload

Overview

File uploads are agent-initiated pulls. The agent requests chunks from Mythic rather than Mythic pushing data. This allows the agent to control throughput and channel selection.

How Files Get to Mythic

When an operator issues a command with a File type parameter:

  1. Operator selects a file in the UI
  2. Mythic uploads and stores the file, assigns a UUID
  3. The command parameters contain the file UUID, not raw bytes
  4. Example parameters sent to agent: {"file": "uuid-here", "path": "/some/path"}

Server-Side: Accessing File Data in create_go_tasking

Get File Contents (Python)

async def create_go_tasking(self, taskData: PTTaskMessageAllData) -> PTTaskCreateTaskingMessageResponse:
    response = PTTaskCreateTaskingMessageResponse(
        TaskID=taskData.Task.ID, Success=True,
    )
    file_resp = await SendMythicRPCFileGetContent(MythicRPCFileGetContentMessage(
        AgentFileId=taskData.args.get_arg("file"),
    ))
    if file_resp.Success:
        # file_resp.Content contains raw bytes
        pass
    else:
        raise Exception("Error from Mythic: " + str(file_resp.error))
    return response

Get File Metadata (Python)

file_resp = await SendMythicRPCFileSearch(MythicRPCFileSearchMessage(
    TaskID=taskData.Task.ID,
    AgentFileID=taskData.args.get_arg("file")
))
if file_resp.Success and len(file_resp.Files) > 0:
    original_filename = file_resp.Files[0].Filename

Swap File UUID for Contents

If your agent needs the raw file bytes in the tasking parameters (rather than pulling chunks):

file_resp = await SendMythicRPCFileGetContent(MythicRPCFileGetContentMessage(
    AgentFileId=taskData.args.get_arg("file"),
))
if file_resp.Success:
    import base64
    taskData.args.add_arg("file", base64.b64encode(file_resp.Content).decode())

Register a New File

To register a file created by the container (not uploaded by the operator):

await SendMythicRPCFileCreate(MythicRPCFileCreateMessage(...))

Agent-Side: Chunked File Pull

The agent requests the file chunk by chunk:

Request Each Chunk

{
    "action": "post_response",
    "responses": [
        {
            "upload": {
                "chunk_size": 512000,
                "file_id": "UUID of the file",
                "chunk_num": 1,
                "full_path": "/full/path/on/target"
            },
            "task_id": "task uuid"
        }
    ]
}

Request Fields

Field Required Description
chunk_size Yes Bytes per chunk requested
file_id Yes UUID of the file to download
chunk_num Yes 1-based. Which chunk to fetch.
full_path Conditional Full path where file will be written on target. Required for Mythic to track the file as written to disk. Omit if file stays in memory only.
task_id Yes Associated task UUID

Response from Mythic

{
    "action": "post_response",
    "responses": [
        {
            "status": "success",
            "total_chunks": 4,
            "chunk_num": 1,
            "chunk_data": "base64 encoded chunk",
            "file_id": "file UUID",
            "task_id": "task UUID"
        }
    ]
}

Response Fields

Field Description
total_chunks Total number of chunks given the requested chunk_size
chunk_num Which chunk Mythic is returning
chunk_data Base64-encoded file data

The agent repeats the request for chunk_num 2 through total_chunks.

Error Response

{
    "action": "post_response",
    "responses": [
        {
            "total_chunks": 0,
            "chunk_num": 0,
            "chunk_data": "",
            "file_id": "",
            "task_id": "",
            "status": "error",
            "error": "error message"
        }
    ]
}

full_path and Disk Tracking

  • If full_path is provided: Mythic records the file as written to disk on the target, visible in the Files search page
  • If full_path is empty or omitted: Mythic treats the transfer as memory-only (no disk artifact recorded)
  • This is separate from artifact reporting on the Artifacts page

File Reuse

Files can be pulled multiple times by default. If delete_after_fetch was set to True on the file, it will be removed from disk after the first complete fetch and cannot be re-used.

Why Not Inline Files in Tasking?

Files are kept separate from tasking JSON to:

  • Keep initial tasking messages small
  • Allow the agent to use a different channel for file transfer (e.g., HTTP for files while using DNS for C2)
  • Support caching and rate limiting on file transfers
  • Prevent exploding DNS or other constrained channels with large file data