generate-personalized-gift
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted external data from LinkedIn profiles and retailer websites using Claygent.
- Ingestion points: Public web content processed by the 'Research Interests' and 'Purchasable Product Link' nodes, as defined in
references/prompts.md. - Boundary markers: There are no explicit delimiters or instructions provided to the AI to ignore instructions potentially embedded within the retrieved web content.
- Capability inventory: The skill is capable of workflow creation and execution through the
clayplugin, and it outputs text to the user conversation or CSV files. It does not have permissions to autonomously purchase or ship items. - Sanitization: The skill lacks mechanisms to sanitize or validate the external content before it is processed by the LLM models.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute various
clayCLI commands (clay whoami,clay workflows list,clay workflows get, etc.) to verify the environment and manage workflow automation. These are standard operations for a skill designed to interact with the author's (clay-run) platform.
Audit Metadata