The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to download and install a companion repository from an untrusted GitHub source.
Evidence: git clone https://github.com/krusemediallc/arcads-claude-code in SKILL.md.
Note: The source repository krusemediallc/arcads-claude-code is not associated with a known trusted vendor or well-known service, posing a supply chain risk.
[INDIRECT_PROMPT_INJECTION]: The skill relies on external prompt templates that could be manipulated to influence the agent's behavior or output.
Ingestion points: The skill reads prompt formulas from prompting/prompt-library/<model>.md within the external companion repository.
Boundary markers: No specific delimiters or safety instructions are mentioned to isolate these external prompt templates from the agent's core logic.
Capability inventory: The skill uses network operations to interact with the Arcads API (https://external-api.arcads.ai) for video and image generation.
Sanitization: There is no evidence of sanitization or validation of the content retrieved from the external prompt library before it is used in API calls.

arcads-broll