news-summary
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external RSS feeds to generate summaries, which creates an attack surface for malicious instructions embedded in the news content.
- Ingestion points: News content is fetched from BBC, Reuters, NPR, and Al Jazeera RSS feeds (SKILL.md).
- Boundary markers: No specific delimiters or instructions are used to separate the external content from the agent's instructions.
- Capability inventory: The skill performs network operations via
curl, text manipulation withsed/grep, and file writes to/tmp/(SKILL.md). - Sanitization: Employs simple HTML tag stripping via
sed, which does not protect against semantic prompt injection. - [COMMAND_EXECUTION]: Executes shell commands including
curl,grep, andsedto retrieve and parse RSS data. These are localized to the news-fetching task. - [EXTERNAL_DOWNLOADS]: Downloads news data from trusted international sources (BBC, Reuters, NPR, Al Jazeera) and uses the OpenAI API for audio generation. These interactions involve reputable service providers.
Audit Metadata