The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external contact information (such as names, emails, and notes) from the Google People API, which could be used to host malicious instructions designed to manipulate agent behavior.\n- Ingestion points: Data is ingested via list, search, and get methods for contactGroups, otherContacts, and people resources as documented in SKILL.md.\n- Boundary markers: The instructions in SKILL.md do not define delimiters or provide specific warnings for the agent to ignore instructions embedded within the retrieved contact data.\n- Capability inventory: The skill enables the agent to modify state through create, update, and delete operations and inspect tool metadata via gws schema as described in SKILL.md.\n- Sanitization: No sanitization or validation logic is specified for the data content retrieved from the external API sources.

gws-people