team-builder
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run the
claude agentsCLI command to retrieve a list of available agents and plugins installed in the environment. - [INDIRECT_PROMPT_INJECTION]: The skill operates as a multi-agent orchestrator that ingests data from external markdown files, creating a surface for indirect prompt injection.
- Ingestion points: The skill reads agent persona definitions from files located in
./agents/**/*.mdand~/.claude/agents/**/*.md(SKILL.md). - Boundary markers: The skill does not employ explicit boundary markers or delimiters to separate the ingested persona content from the user-provided task instructions.
- Capability inventory: The skill is capable of invoking the
Agenttool in parallel, passing the raw content of the markdown files directly into the sub-agent's prompt. - Sanitization: There is no evidence of content sanitization, filtering, or input validation performed on the agent personas before they are used to influence sub-agent behavior.
Audit Metadata