The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for ingesting and processing data from external Content Management Systems, which creates a surface for indirect prompt injection attacks.\n
Ingestion points: External configuration data and content retrieved from WordPress, Strapi, Contentful, or Ghost platforms during connection testing and cross-referencing in SKILL.md.\n
Boundary markers: The skill does not provide specific delimiters or instructions to ignore embedded commands within the external data.\n
Capability inventory: The skill is granted access to the Bash and Write tools in SKILL.md, providing a mechanism for potential exploitation if malicious instructions are processed.\n
Sanitization: There is no evidence of validation or sanitization logic for the data ingested from the CMS platforms in the analyzed file.

setup-cms