ql-execute

Warn

Audited by Gen Agent Trust Hub on Mar 25, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on the execution of a local shell script and system build tools to perform its tasks.
  • Evidence: Instructs the agent to run ./quantum-loop.sh with various flags and mentions the availability of npm, pip, and cargo as prerequisites.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it incorporates instructions and data from external project files into the agent's decision-making process.
  • Ingestion points: Reads state and dependency data from quantum.json and execution instructions from agents/orchestrator.md (SKILL.md).
  • Boundary markers: Absent; there are no specific markers or instructions to treat the content of the ingested files as untrusted data rather than direct commands.
  • Capability inventory: Executes local shell scripts, performs git operations, and runs system build tools (SKILL.md).
  • Sanitization: Absent; the skill does not specify any validation or sanitization of the logic retrieved from the project files.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 25, 2026, 02:58 AM
Security Audit — agent-trust-hub — ql-execute