team-builder
Pass
Audited by Gen Agent Trust Hub on Mar 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted markdown files from the local filesystem and global configuration directories and uses their raw content to define sub-agent behavior.
- Ingestion points: SKILL.md (Step 4) details reading files from
./agents/and~/.claude/agents/. - Boundary markers: Absent. The prompt template
"{agent file content}\n\nTask: {task description}"does not use XML tags, delimiters, or explicit instructions to prevent the file content from overriding the user's task description. - Capability inventory: SKILL.md (Step 4) invokes the
Agenttool to execute the final combined prompt. - Sanitization: Absent. There is no evidence of escaping, validation, or filtering of the content read from the external markdown files before interpolation into the prompt.
Audit Metadata