handoff
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to interact with the system clipboard.
- Evidence: The command
echo "PROMPT_CONTENT" | pbcopyis used to store the generated summary on macOS systems. - [INDIRECT_PROMPT_INJECTION]: The skill processes the entire current conversation history to generate a summary, creating a surface for indirect prompt injection.
- Ingestion points: The agent's current session context, which includes all previous user inputs and tool outputs.
- Boundary markers: Absent; the skill does not wrap the summarized content in protective delimiters or include instructions to ignore embedded commands within the summary.
- Capability inventory: Shell command execution via
pbcopy. - Sanitization: Absent; the skill performs a summarization of existing content without filtering for potentially malicious instructions that might be carried over into the new session's bootstrap prompt.
Audit Metadata