team-builder
Pass
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via malicious markdown files.
- Ingestion points: Processes markdown files from project-local directories and global directories like
~/.claude/agents/(Step 1). - Boundary markers: No delimiters or warnings are used when reading file content; the raw content is concatenated directly into the prompt (Step 4).
- Capability inventory: Spawns sub-agents via the
Agenttool using the read content as the primary instruction set (Step 4). - Sanitization: No sanitization, validation, or escaping of the markdown content is performed before interpolation into the sub-agent prompt.
Audit Metadata