The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill documents the use of moda canvas edit and the markup tag, which allow for the execution of synchronous JavaScript within a platform-enforced sandbox. This environment includes significant safety restrictions, such as blocking eval, require, and network-related globals, while enforcing execution time and code size limits.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for ingesting external data from files and existing canvas states to automate design tasks. 1. Ingestion points: moda canvas read and moda file show are used to extract content for layout generation. 2. Boundary markers: Explicit delimiters or ignore instructions for third-party data are not defined in the skill body. 3. Capability inventory: The agent uses moda canvas markup and moda canvas edit to process this data. 4. Sanitization: Data is processed through a specialized design toolchain with built-in structural constraints and a restricted sandbox.
[COMMAND_EXECUTION]: The skill operates via a restricted set of vendor-specific commands. The allowed-tools frontmatter field restricts the environment to Bash(moda:*) and Read operations, ensuring that the agent can only interact with the intended design platform services.

moda-social