skills/davidondrej/skills/rename/Gen Agent Trust Hub

rename

Pass

Audited by Gen Agent Trust Hub on Oct 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the room-cli command-line utility to perform thread metadata updates. Specifically, it executes room-cli thread update --self --title "[new name]" to change the session title.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for indirect prompt injection as it processes the content of the current thread to generate a summary title.
  • Ingestion points: The agent analyzes the entire thread or session history to determine the "main thing happening" (SKILL.md).
  • Boundary markers: The prompt provides a strict constraint for the model to pick "2-5 clear lowercase words", which limits the potential for complex instruction injection into the title.
  • Capability inventory: The skill possesses the capability to modify thread metadata using the room-cli tool (SKILL.md).
  • Sanitization: The generated title is interpolated into a bash command within double quotes, providing basic shell escaping for the generated string.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 20, 2026, 06:48 AM