The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests data from external Google Slides presentations, which can contain instructions that attempt to influence agent behavior.
Ingestion points: Data enters the context via read-slide, list-slides, and raw commands as documented in SKILL.md.
Boundary markers: The skill suggests using the --wrap-untrusted flag to assist the agent in isolating external content in SKILL.md.
Capability inventory: The skill possesses broad manipulation permissions, including slide deletion, text replacement, and presentation cloning in SKILL.md.
Sanitization: Safety is primarily handled through the tool-level --wrap-untrusted flag.
[EXTERNAL_DOWNLOADS]: The skill facilitates the retrieval of remote assets for presentation use.
Evidence: The insert-image and replace-slide commands explicitly support fetching content from public URLs in SKILL.md.
Evidence: The export and thumbnail commands involve downloading files to the local environment in SKILL.md.

gog-slides