agenttrace-session-audit

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute the agenttrace command-line utility and go run ./cmd/agenttrace to perform session audits, health checks, and comparisons.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes session logs and trace data from external coding agents, which serves as a potential data ingestion surface for untrusted content.
  • Ingestion points: The agent reads session logs from tools like Claude Code, Cursor, and Aider via files (agenttrace path/to/session.json) or directory scanning (agenttrace --overview -d path).
  • Boundary markers: There are no explicit instructions for the agent to use delimiters or ignore instructions embedded within the logs it is auditing.
  • Capability inventory: The skill allows shell command execution (agenttrace) and file writing (-o agenttrace-overview.md).
  • Sanitization: The instructions include a manual safety note to avoid printing secrets found in logs, but no automated sanitization or filtering logic is defined for the log ingestion process.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 09:29 AM
Security Audit — agent-trust-hub — agenttrace-session-audit