The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from sources such as emails, hotel communications, and airline profiles to extract booking details and traveler information, which creates an attack surface for indirect prompt injection. * Ingestion points: The skill reads content from 'relevant emails', 'hotel pickup emails', 'airline profiles', and 'entry forms' as described in SKILL.md. * Boundary markers: The instructions do not specify technical delimiters or 'ignore instructions' markers for the ingested data, although they do reference a confirmation_policy and standing_authorization_for_information_disclosure to govern interactions. * Capability inventory: The skill possesses the capability to perform flight bookings, execute airline check-ins, and prepare communication drafts as specified in SKILL.md. * Sanitization: The instructions include behavioral sanitization, such as forbidding the repetition of sensitive details like full date of birth in chat and requiring verification of information from multiple sources before use.

flights