wave-executor
Pass
Audited by Gen Agent Trust Hub on Mar 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local orchestration tool at '.claude/tools/cli/wave-executor.mjs' using Node.js to manage batch waves. This execution is performed securely using 'spawnSync' with shell set to false to prevent shell injection.
- [EXTERNAL_DOWNLOADS]: The skill references and requires the '@anthropic-ai/claude-agent-sdk' package, which is an official library provided by a trusted organization.
- [PROMPT_INJECTION]: The skill ingests task definitions and templates from a user-provided JSON plan file. Ingestion points: Data is read from the file path specified in the '--plan' argument. Boundary markers: No explicit delimiters are used in the provided scripts to isolate external plan content. Capability inventory: The skill utilizes tools such as Bash, Write, Glob, and Grep. Sanitization: The pre-execute hook performs JSON schema validation on the plan file to ensure it contains a valid array of waves before processing begins.
Audit Metadata