The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references external resources, including documentation at composio.dev and an MCP server endpoint at https://rube.app/mcp. These are official service components belonging to the toolkit provider and are necessary for the skill's primary functionality.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute tools such as RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH. These tools are part of the intended Adobe automation workflow and are used to perform tasks within the scope of the integrated service.
[PROMPT_INJECTION]: The skill ingests dynamic tool schemas and output data from the RUBE_SEARCH_TOOLS command, representing a standard surface for indirect prompt injection common in tool-integrated skills.
Ingestion points: Dynamic tool schemas and usage guidelines are fetched from RUBE_SEARCH_TOOLS at runtime and incorporated into the agent's workflow (SKILL.md).
Boundary markers: The skill does not define specific delimiters or "ignore instructions" directives for the data retrieved from tool commands.
Capability inventory: The skill has the capability to execute complex Adobe workflows and interact with a remote workbench through RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH.
Sanitization: No explicit data sanitization or schema validation is performed within the skill instructions; it relies on the agent's internal safety guardrails and the provider's API integrity.

adobe-automation