qrspi-create-design-discussion
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing external data to generate architectural decisions and documents.\n
- Ingestion points: The skill reads research documents, original tickets, and answered questions from the
thoughts/shared/directory (SKILL.md, Step 1).\n - Boundary markers: Absent; the instructions do not specify any delimiters (e.g., XML tags or triple backticks) to isolate untrusted file content from the agent's primary instructions.\n
- Capability inventory: The skill writes Markdown design documents, generates HTML slide decks using reveal.js, and spawns a separate
codebase-pattern-finderagent.\n - Sanitization: Absent; the instructions do not provide requirements for sanitizing or escaping the content of research documents or tickets before they are synthesized into the design documents or HTML output.\n- [DATA_EXPOSURE]: The skill reads from and writes to the local
thoughts/shared/directory. This file system access is scoped to the project environment and is necessary for the intended design-tracking functionality.\n- [COMMAND_EXECUTION]: The skill utilizes an agent orchestration pattern to spawn a 'codebase-pattern-finder'. This is an internal delegation of tasks to another agent and does not involve the execution of arbitrary shell commands.
Audit Metadata