cli-anything-calibre

Warn

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the package directly from a remote GitHub repository (github.com/HKUDS/CLI-Anything.git) using pip, which bypasses official package registries.
  • [COMMAND_EXECUTION]: The harness functions by executing system-level binaries such as 'calibredb', 'ebook-convert', and 'ebook-meta', passing arguments to perform file system and library modifications.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to read and process various e-book formats (EPUB, PDF, MOBI, etc.) from external sources; these files can contain hidden instructions that may attempt to override the AI agent's behavior during content extraction or metadata analysis.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 15, 2026, 05:21 PM
Security Audit — agent-trust-hub — cli-anything-calibre