vm-trace-analyzer

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted JSON trace files.
  • Ingestion points: User-provided trace JSON is read in SKILL.md and processed via scripts/parse_trace.py.
  • Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential instructions embedded in trace message fields.
  • Capability inventory: The skill utilizes subprocess calls to execute local analysis scripts.
  • Sanitization: No sanitization is performed on the trace data before the agent interprets it for report generation.
  • [EXTERNAL_DOWNLOADS]: The skill fetches official VictoriaMetrics changelogs from GitHub.
  • The scripts/check_changelog.py script downloads markdown files from the official VictoriaMetrics repository on github.com. This aligns with the skill's purpose and uses a well-known, trusted service.
  • [COMMAND_EXECUTION]: The skill executes local Python scripts to analyze data.
  • SKILL.md contains instructions to run scripts/parse_trace.py and scripts/check_changelog.py to extract performance metrics and version information.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 07:24 AM