The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external Google Slides presentations, creating a surface for potential indirect instructions.
Ingestion points: The presentations.get method in SKILL.md allows the agent to retrieve and process the content of existing presentations.
Boundary markers: The skill documentation does not provide specific markers or instructions to delimit or isolate content retrieved from presentations.
Capability inventory: The skill possesses write capabilities through presentations.create and presentations.batchUpdate, which could be abused if the agent follows instructions embedded in a slide.
Sanitization: There is no evidence of content sanitization or validation for the data returned by the API calls.
[COMMAND_EXECUTION]: The skill relies on the execution of the gws command-line tool to perform its operations.
Evidence: The skill uses gws slides for API interactions and gws schema for discovering command parameters and structures.

gws-slides