weather

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION] (LOW): The skill allows the use of curl or wget to fetch raw content when web search tools fail. While the primary intent is fetching "public content," this capability creates a surface for Server-Side Request Forgery (SSRF) if the agent is directed to request internal or malicious URLs.
  • [PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection (Category 8c) because it reads and processes untrusted data from external websites.
  • Ingestion points: External weather data and web pages retrieved via web search tools or curl/wget as specified in SKILL.md.
  • Boundary markers: Absent; the instructions do not specify the use of delimiters or warnings to ignore instructions embedded within the fetched data.
  • Capability inventory: Command execution (curl, wget), web search, and web reading across the toolset.
  • Sanitization: Absent; the skill lacks explicit logic to sanitize or filter the content of the retrieved weather reports.
  • [EXTERNAL_DOWNLOADS] (LOW): The skill references and downloads data from external domains. Several of these are recognized authoritative weather sources (e.g., weather.gov, nmc.cn), which qualifies for a severity downgrade under the [TRUST-SCOPE-RULE] for those specific references.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:09 PM