workflow-execute

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill's behavior aligns with its stated purpose of executing development plans. No malicious patterns were identified.\n- [COMMAND_EXECUTION]: The skill facilitates shell command execution specifically for running tests and verifying code implementation, which are standard project operations.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it reads and follows instructions from local project files such as plan.md and analysis.md. This is a requirement for the skill's functionality.\n
  • Ingestion points: Reads content from docs/work/*/plan.md, analysis.md, and the __prototype__/ directory.\n
  • Boundary markers: Absent; the skill does not use specific delimiters or instructions to ignore embedded commands within the ingested files.\n
  • Capability inventory: Includes file creation/modification, shell command execution (via test running), and software dependency installation.\n
  • Sanitization: No validation or sanitization is performed on the content of the plan files before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 09:22 AM
Security Audit — agent-trust-hub — workflow-execute