The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the meitu CLI to interface with Meitu's visual processing models for tasks such as image generation, editing, and video creation.
[EXTERNAL_DOWNLOADS]: The documentation guides users to install the meitu-cli package via npm, which is a tool required for the skill's core functionality.
[DATA_EXFILTRATION]: User-provided images and generated prompts are transmitted to the Meitu OpenAPI for processing. This data flow is central to the skill's purpose and is clearly disclosed to the user in the data disclaimer section.
[PROMPT_INJECTION]: The skill ingests local user data (e.g., MEMORY.md, PROFILE.md) to personalize generated content. While this creates a potential surface for indirect prompt injection from local files, the risk is minimal as the input is used for prompt augmentation within structured templates.
Ingestion points: Reads personalization data from files including USER.md, MEMORY.md, and PROFILE.md located in the workspace.
Boundary markers: The agent uses predefined formulas and assembly rules to wrap user context into narrative prompts, though it lacks explicit 'ignore instructions' delimiters for the ingested data.
Capability inventory: The skill utilizes the meitu CLI for all visual processing tasks, including image-generate, image-edit, and face-swap.
Sanitization: Content is processed and interpreted by the agent before being passed to the CLI as part of a structured prompt.
[CREDENTIALS_UNSAFE]: The skill manages API access keys through environment variables or a local configuration file (~/.meitu/credentials.json), adhering to standard credential management practices for developer tools.

meitu-visual-me