wave-executor

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes a local CLI tool located at .claude/tools/cli/wave-executor.mjs using the child_process.spawnSync method with shell: false. This approach is a security best practice that prevents command injection vulnerabilities by ensuring that arguments are not interpreted by a shell.
  • [PROMPT_INJECTION]: The skill consumes user-provided JSON plan files that contain a promptTemplate field, which is used to instruct sub-agents. This creates a surface for indirect prompt injection. 1. Ingestion points: The plan JSON file path provided as an argument. 2. Boundary markers: No explicit delimiters or 'ignore' instructions for the template content were found in the provided scripts. 3. Capability inventory: The skill has the capability to orchestrate multiple sub-agents with potentially broad tool access. 4. Sanitization: The content is parsed as JSON but does not undergo specific sanitization for prompt instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 12:54 AM
Security Audit — agent-trust-hub — wave-executor