fountain-daily-growth
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The trend-discovery module (modules/trend-discovery/MODULE.md) fetches headlines and summaries from Google News RSS. This external, untrusted content is used to create briefs that influence the behavior of downstream skills, creating a potential surface for prompt injection via manipulated news content.
- Ingestion points: News headlines and snippets fetched via RSS in trend-discovery.
- Boundary markers: The skill does not explicitly define delimiters to separate external news content from instructions in the generated briefs.
- Capability inventory: The skill can execute network requests, write to project preferences, and invoke other specialized skills.
- Sanitization: There is no mention of filtering or sanitizing news content before it is processed by the agent.
- [DYNAMIC_EXECUTION]: Instructions in HOUSEKEEPING.md permit the agent to generate and execute temporary scripts to handle batch API requests more efficiently. While this is restricted to throwaway tasks, it involves runtime code generation and execution.
- [EXTERNAL_DOWNLOADS]: The skill uses curl to download RSS feeds from Google News (news.google.com). This is a well-known service and the operation is fundamental to the skill's discovery purpose.
- [COMMAND_EXECUTION]: The skill requires the use of curl for performing network requests to retrieve external data and interacts with the Fountain Social and Project APIs to manage state and content.
Audit Metadata