meta-cognition-parallel

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a vulnerability to indirect prompt injection (Category 8) due to its handling of external data.
  • Ingestion points: The skill ingests untrusted user data via the $ARGUMENTS variable in SKILL.md.
  • Boundary markers: The skill uses simple markdown headers (e.g., ## User Query) to separate instructions from user input. These are weak delimiters that can be bypassed by an adversary providing crafted input designed to hijack the sub-agent's behavior.
  • Capability inventory: The skill has the capability to read local files from the ../../agents/ directory and execute sub-agents using the Task() orchestration function.
  • Sanitization: There is no evidence of input validation, escaping, or filtering for the $ARGUMENTS content before it is passed to the parallel sub-agents.
  • [COMMAND_EXECUTION]: The skill utilizes a Task() function to launch parallel processes. In this context, these are framework-level sub-agents rather than arbitrary shell commands, representing the intended orchestration logic of the vendor.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 11:06 PM