The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, creating a surface for instructions embedded in customer feedback to influence agent behavior.
Ingestion points: Data is collected from support tickets, chat logs, and SNS mentions as defined in the VOC Analysis workflow in SKILL.md and the Data Collector role in references/feedback-analyzer.md.
Boundary markers: The instructions do not specify the use of delimiters (e.g., triple quotes or XML tags) or explicit instructions for the agent to ignore commands within the analyzed text.
Capability inventory: The skill generates response scripts and reports, writes files to a _workspace/ directory, and passes data to downstream skills such as moai-writer and moai-officer.
Sanitization: While the skill requires masking of personally identifiable information (PII) in output reports, it does not define sanitization or validation logic for the input text to mitigate injection risks.

cs-escalation