NYC
skills/boshu2/agentops/converter/Gen Agent Trust Hub

converter

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONNO_CODE
Full Analysis
  • COMMAND_EXECUTION (LOW): The skill documentation instructs the agent to execute a local shell script bash skills/converter/scripts/convert.sh. This is an expected pattern for CLI-based tools but requires that the script itself be audited for safe parameter handling.\n- FILE_SYSTEM_OPERATIONS (LOW): The skill performs destructive write operations by deleting the target output directory before writing new content. This poses a risk of unintended data loss if the output path is incorrectly specified by a user or agent.\n- NO_CODE (LOW): The implementation logic for the converter is contained within scripts/convert.sh, which was not included in the provided file set. This prevents verification of the script's safety and integrity.\n- INDIRECT_PROMPT_INJECTION (LOW): The skill is designed to parse untrusted external directories.\n
  • Ingestion points: The skill-dir directory, including its SKILL.md, references/, and scripts/ subdirectories.\n
  • Boundary markers: None identified; the skill parses content based on YAML and Markdown delimiters.\n
  • Capability inventory: File system read/write and local bash script execution.\n
  • Sanitization: None documented; the skill assumes valid AgentOps skill formats.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 08:55 PM