The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing the canva command-line utility to perform design operations, asset management, and authentication tasks.
[EXTERNAL_DOWNLOADS]: The skill uses curl to fetch thumbnail images from Canva's infrastructure for display as previews within the agent's workspace.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes user-controlled content from Canva designs and presenter notes, which could contain malicious instructions designed to manipulate the agent's behavior.
Ingestion points: Reads design metadata, layout text, and presenter notes via read-design and get-design-content (SKILL.md).
Boundary markers: The instructions do not specify the use of clear delimiters or warnings to separate design content from the agent's system instructions.
Capability inventory: The agent can perform write and delete operations on designs, upload local files, and execute CLI commands (manifest.yaml, SKILL.md).
Sanitization: There is no evidence of sanitization or filtering applied to ingested design content before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill employs a call-tool command that dynamically executes different MCP tools based on names and JSON arguments provided at runtime.
[DATA_EXFILTRATION]: The skill includes an upload-file command that takes a file path as an argument to read and upload local filesystem data to Canva. This capability is managed through the manifest's requirement for explicit user approval for imports.

canva