qa-personality-builder
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill dynamically generates and writes new
SKILL.mdfiles to the.claude/skills/directory. These files represent new agent personalities with specific instructions and tool access, effectively allowing the skill to create new, persistent agent capabilities at runtime. - [INDIRECT_PROMPT_INJECTION]: The builder workflow creates a vulnerability surface by interpolating user-provided text into new agent instructions without explicit sanitization.
- Ingestion points: User responses to questions regarding testing specialties and specific test scenarios in
SKILL.md(Mode A). - Boundary markers: Absent; user input is placed directly into the generated skill sections without delimiters or instructions to ignore embedded commands.
- Capability inventory: Generated skills can be assigned powerful tools including
Bash,WebFetch,Read, andPlaywrightbrowser automation. - Sanitization: None; the builder does not provide logic for escaping or validating the user's input before it is written into the new skill's instruction set.
Audit Metadata