cook-auto-parallel

Pass

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes external input and moves to autonomous execution.- Ingestion points: Untrusted data enters the agent context through the $ARGUMENTS variable in SKILL.md and by reading various project files such as docs/project-reference/ and .claude/skills/shared/.- Boundary markers: The skill uses <tasks> tags to wrap input arguments, which provides some structural separation but does not include explicit instructions to ignore embedded commands.- Capability inventory: The skill coordinates several subagents with implementation capabilities, including file system modifications, test execution, and git management (via fullstack-developer and git-manager agents).- Sanitization: No validation or sanitization of the input arguments or the content of the read files is performed before they are used to generate an execution plan.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 21, 2026, 03:49 AM