The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes potentially untrusted external data and user source materials, which creates a surface for indirect prompt injection attacks.
Ingestion points: The skill is instructed to read "user's chosen stack" and "source material" in SKILL.md, and fetches external media or data using sandbase_run and sandbase_run_get as described in references/sandbase-api-map.md.
Boundary markers: The instructions do not specify the use of delimiters or specific system instructions to ignore embedded commands within the ingested content.
Capability inventory: The skill possesses the capability to execute tools (sandbase_run) and generate functional artifacts (code implementation and media verification).
Sanitization: There are no explicit requirements in the workflow to sanitize, validate, or escape external data before it is incorporated into the implementation or output artifacts.

gsap-plugins