press-release-writing

Fail

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides installation instructions that use the curl -fsSL https://cli.inference.sh | sh pattern. This method downloads and executes a script directly from a remote server in the user's shell environment without prior inspection or verification.
  • [EXTERNAL_DOWNLOADS]: Setup and operation of the skill involve downloading binaries from dist.inference.sh. These external resources are required for the tool's core functionality but are retrieved from a non-standard domain.
  • [COMMAND_EXECUTION]: The skill is configured with allowed-tools: Bash(infsh *), which permits the agent to execute shell commands. This capability is used to interact with the infsh CLI for authentication and running search applications.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests untrusted data from the live web via tavily/search-assistant and exa/search tools.
  • Ingestion points: Data enters the context through research commands in SKILL.md using the infsh tool.
  • Boundary markers: No clear delimiters or warnings are used to instruct the agent to ignore instructions embedded within the retrieved search results.
  • Capability inventory: The skill possesses the Bash capability to execute CLI commands and interact with external APIs.
  • Sanitization: There is no evidence of sanitization or filtering of the content returned from external search providers before it is processed by the agent.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 4, 2026, 12:29 PM