content-repurposing

Fail

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill directs users to execute curl -fsSL https://cli.inference.sh | sh. This pattern is extremely dangerous as it executes unverified code from an untrusted remote source with the privileges of the local shell.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes npx skills add inference-sh/skills@... to fetch and install external components from a non-trusted repository, which can lead to supply chain attacks.
  • [COMMAND_EXECUTION]: The skill defines allowed-tools: Bash(infsh *), granting the agent unrestricted access to the infsh CLI tool and its various remote application capabilities.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection when processing external content like podcast transcripts and blog posts.
  • Ingestion points: STT app output and blog content (SKILL.md).
  • Boundary markers: No delimiters or safety instructions are used to isolate ingested data.
  • Capability inventory: Capabilities include social media posting (x/post-create) and video generation (veo-3-1-fast).
  • Sanitization: No validation or sanitization is performed on the data 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
CRITICAL
Analyzed
Mar 9, 2026, 10:14 PM