The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to download a script from a remote URL and pipe it directly to the shell (curl ... | sh) if the Higgsfield CLI is missing. This allows the execution of arbitrary, unverified code from an external source at runtime.
[EXTERNAL_DOWNLOADS]: The skill depends on a CLI installer hosted at raw.githubusercontent.com under the higgsfield-ai organization. This external resource is not part of the verified vendor infrastructure for this environment.
[COMMAND_EXECUTION]: The skill uses the Bash tool to perform system checks, install the Higgsfield CLI, and execute commands for image generation and account authentication.
[PROMPT_INJECTION]: The skill ingests data from local configuration files which could contain instructions that influence the agent's behavior.
Ingestion points: The agent is instructed to read brand/voice-profile.md, brand/visual-style.md, and brand/creative-kit.md upon activation.
Boundary markers: No delimiters or warnings are provided to the agent to treat the content of these files as data rather than instructions.
Capability inventory: The agent has access to Bash (shell execution), Read, and Write tools, providing a broad capability set if the agent is influenced by injected content.
Sanitization: There is no evidence of validation or filtering for the content read from the brand files.

higgsfield-product-photoshoot