figure-it-out

Warn

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill implements logic to find a specific validation script (check-omc-playbooks.mjs) by searching directory paths relative to the omc binary and plugin environment variables. This script is then executed using node.
  • [COMMAND_EXECUTION]: Shell commands are used to discover the environment and resolve file paths, including the use of command -v, dirname, and pwd -P.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to generate new instruction artifacts (playbooks) that the agent will follow in future interactions, establishing a surface for indirect instruction influence.
  • Ingestion points: The skill writes Markdown files to a dynamically determined playbooks_dir and reads from project script directories.
  • Boundary markers: Generated playbook files use YAML frontmatter to separate configuration metadata from the instruction body.
  • Capability inventory: The skill possesses file-write permissions and the ability to execute JavaScript via a local node environment.
  • Sanitization: The skill relies on an external local script for validation but does not detail how the content of the generated playbooks is sanitized or escaped to prevent malicious instruction embedding.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Nov 8, 2026, 07:21 PM