daily-gift

Warn

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: MEDIUMPERSISTENCECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PERSISTENCE]: The skill configures a recurring cron job for autonomous daily gift delivery, enabling background code execution outside of active user sessions.
  • [COMMAND_EXECUTION]: Shell scripts are used to handle the gift rendering pipeline and asset management, which involves direct execution of underlying system commands.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes shell scripts to perform network requests to external APIs for image and video generation and to fetch creative assets from remote hosting services.
  • [DATA_EXFILTRATION]: By reading conversation history and memory files and then sending data to external rendering APIs, the skill creates a surface where sensitive personal information could be sent to third-party services.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests broad conversation context to drive its editorial judgment, which could allow malicious instructions in user data to override the agent's behavior.
  • Ingestion points: Reads conversation context and memory files from the local workspace.
  • Boundary markers: No documentation is provided regarding the use of delimiters or instruction-ignoring markers for ingested context.
  • Capability inventory: The skill can execute shell scripts, make network calls, and write to the local file system.
  • Sanitization: There is no mention of sanitizing or validating the ingested content before it enters the creative pipeline.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 6, 2026, 01:18 PM
Security Audit — agent-trust-hub — daily-gift