klipper-ops

Pass

Audited by Gen Agent Trust Hub on Oct 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses qdbus commands to interact with the Klipper clipboard manager on KDE Plasma systems. This enables the agent to retrieve current clipboard contents and the full clipboard history. Since the clipboard is a common location for sensitive data like passwords, tokens, and personal information, this capability exposes high-value data to the agent context.
  • [INDIRECT_PROMPT_INJECTION]: The skill introduces a surface for indirect prompt injection by reading untrusted data from the system clipboard.
  • Ingestion points: Untrusted content is ingested from the clipboard via getClipboardContents and getClipboardHistoryItem commands documented in SKILL.md.
  • Boundary markers: There are no instructions or markers defined to prevent the agent from following instructions that might be contained within the copied clipboard text.
  • Capability inventory: The skill possesses the ability to read, write, and clear clipboard data using qdbus calls.
  • Sanitization: The skill does not implement any sanitization or filtering of the retrieved clipboard data before it is presented to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 17, 2026, 10:44 PM