session-recorder

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill's instructions require the agent to execute a Python script while passing the entire unedited session history as a command-line argument. This presents a risk of command injection if the session content contains shell-active characters (such as backticks, semicolons, or dollar signs) and the underlying shell execution environment does not properly escape these characters.
  • [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by ingesting and storing untrusted data from the session history.
  • Ingestion points: The <full_unedited_session_content> parameter defined in the command section of SKILL.md.
  • Boundary markers: No boundary markers or delimiters are used to encapsulate the content within the CLI command or the resulting log file.
  • Capability inventory: The skill performs local file system writes via scripts/record_session.py to the user's home directory (~/daily_work).
  • Sanitization: No sanitization, filtering, or escaping of the session content is performed before it is passed to the script or appended to the markdown log file.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 08:42 AM