agent-creator

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local shell scripts ./scripts/init-agent.sh and ./scripts/validate-agent.sh to initialize and validate agent structures during the workflow.
  • [PROMPT_INJECTION]: The skill provides an indirect prompt injection surface by ingesting and processing untrusted user input for agent configuration.
  • Ingestion points: SKILL.md (Phases 1 and 3) where user-provided agent names, missions, and roles are gathered.
  • Boundary markers: Absent in SKILL.md. The input is processed without explicit delimiters or instructions to ignore embedded commands.
  • Capability inventory: SKILL.md (Phases 5 and 6) includes shell command execution and file system write operations for creating agent scaffolds.
  • Sanitization: Absent in SKILL.md. There is no evidence of validation or escaping for the user-provided strings before they are used in file generation or script arguments.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 11:39 AM
Security Audit — agent-trust-hub — agent-creator