daily-news-report
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's core functionality involves ingesting data from external, third-party websites to create summaries, which exposes it to potential indirect prompt injection attacks where malicious instructions hidden in web content could influence the agent.
- Ingestion points: Data enters the agent's context from multiple external URLs (defined in
sources.json) using theWebFetchandmcp__chrome-devtoolstools. - Boundary markers: The skill attempts to mitigate risks by instructing sub-agents to follow a strict JSON schema and output format, which provides some structural separation between the untrusted data and the agent's control flow.
- Capability inventory: The skill has the ability to read and write local files (using
Read,Write, and restrictedBashcommands) and execute tasks via sub-agents using theTasktool. - Sanitization: There is no evidence of pre-processing or sanitization of the scraped HTML or text content to remove potentially malicious hidden instructions before the data is analyzed by the LLM.
Audit Metadata