skills/xenodium/emacs-skills/dired/Gen Agent Trust Hub

dired

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes emacsclient --eval via the Bash tool to send Lisp commands to a running Emacs process.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes file information from the conversation history.
  • Ingestion points: Directory paths and file names from the conversation are interpolated into a Lisp command string.
  • Boundary markers: No delimiters or specific instructions are provided to isolate file data from the Lisp code.
  • Capability inventory: The emacsclient --eval command has full access to the Emacs Lisp interpreter, which can execute system commands.
  • Sanitization: No sanitization or escaping of file names is performed, which could allow a maliciously named file (e.g., one containing quotes or parentheses) to execute unintended Lisp code.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 05:50 AM