export-session-log
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and summarize conversation history, tool results, and file contents. This ingestion of untrusted data creates a surface for indirect prompt injection, where malicious instructions embedded in the processed text could attempt to influence the agent's log generation or subsequent actions.\n
- Ingestion points: Conversation history, tool results, and changed files (as specified in the Workflow section of
SKILL.md).\n - Boundary markers: The instructions do not specify the use of delimiters or unique identifiers to isolate untrusted session data from the agent's primary instructions.\n
- Capability inventory: The skill utilizes local file system writes to save the logs and executes PowerShell commands to retrieve the system time.\n
- Sanitization: The skill provides a negative constraint for the agent to 'Remove secrets, tokens, passwords, personal data, and sensitive local values,' which relies on the LLM's adherence to instructions rather than technical filtering.\n- [COMMAND_EXECUTION]: The skill instructs the agent to execute a PowerShell command to synchronize the log's timestamp with the local system time.\n
- Evidence: The workflow in
SKILL.mdexplicitly directs the agent to runGet-Date -Format \"yyyy-MM-ddTHH:mm:ss\".
Audit Metadata