qa-plan
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from project files and has the capability to modify existing files, creating a potential for indirect prompt injection.
- Ingestion points: Reads all story files (
production/epics/*/story-*.md), referenced GDD files (design/gdd/*.md), the systems index, and the architecture control manifest. - Boundary markers: Absent. The skill instructions do not specify any delimiters or ignore-instructions warnings when interpolating content from these files into the generated QA plan.
- Capability inventory: Includes
Read,Write,Glob,Grep, andmcp__backlog__task_list. It can create new plan files and back-fill existing story files with generated content. - Sanitization: Absent. The skill extracts raw text (acceptance criteria, formulas, edge cases) and incorporates it directly into the generated markdown output.
- [SAFE]: The skill implements a secure practice by using the
AskUserQuestiontool to require explicit user approval before performing any file write operations in Phase 5.
Audit Metadata