debug-training-logs

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes worker and AIStore logs, which are untrusted external inputs, creating a surface for indirect prompt injection. \n
  • Ingestion points: Worker logs ($ARGUMENTS[0]) and AIStore logs ($ARGUMENTS[1]). \n
  • Boundary markers: Absent. The skill does not define delimiters to separate log content from instructions. \n
  • Capability inventory: Access to Bash, Read, Grep, Glob, and Agent tools across multiple scripts enables command execution. \n
  • Sanitization: Absent. Raw log data is parsed for patterns without sanitization.\n- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download diagnostic logs from an AIStore cluster using the ais CLI and suggests the option to skip TLS certificate verification (skip_verify_crt=true).\n- [COMMAND_EXECUTION]: The skill uses the Bash tool to perform log parsing, archive extraction, and diagnostic tool invocation (e.g., grep, tar, sort, ais).
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 07:47 AM
Security Audit — agent-trust-hub — debug-training-logs