press-release-writing
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- Remote Code Execution (CRITICAL): The skill includes the instruction
curl -fsSL https://cli.inference.sh | sh. This is a classic untrusted remote code execution (RCE) pattern that downloads and executes an opaque script from a non-whitelisted domain directly in the user's shell. - Indirect Prompt Injection (HIGH):
- Ingestion points: The skill ingests untrusted external data via
infsh app run tavily/search-assistantandexa/search(SKILL.md). - Boundary markers: Absent. There are no delimiters or instructions to treat the search output as untrusted data.
- Capability inventory: The skill has
allowed-tools: Bash(infsh *), providing the ability to execute commands that could be influenced by malicious content returned from search results. - Sanitization: Absent. External content is used directly to "fact-check" or provide "context" without filtering.
- External Downloads (MEDIUM): The skill references external packages via
npx skills add inferencesh/skills. Since theinferenceshorganization is not on the Trusted External Sources list, these dependencies are considered unverifiable and pose a supply-chain risk.
Recommendations
- CRITICAL: Downloads and executes remote code from untrusted source(s): https://cli.inference.sh - DO NOT USE
- AI detected serious security threats
Audit Metadata