agent-role-scaffolder
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill serves as a generator for agent role definitions, which introduces an attack surface for indirect prompt injection. Malicious instructions provided by a user during the scaffolding process could be embedded into the resulting agent's permanent system instructions.
- Ingestion points: User-provided answers to the scaffolding questionnaire in
references/subagent-role-scaffold.mdandreferences/companion-worker-scaffold.md, specifically related to the agent's purpose, triggers, and task instructions. - Boundary markers: The skill does not explicitly instruct the agent to use delimiters or ignore potentially injected instructions within the user-provided descriptions.
- Capability inventory: The skill allows for the creation of durable role definitions (
~/.claude/agents/<name>.md) and temporary worker instances, giving the agent the ability to write to the file system and launch sub-agents. - Sanitization: There is no evidence of logic for sanitizing, escaping, or validating the input provided by the user before it is formalized into the new agent's role definition.
Audit Metadata