rdq
Pass
Audited by Gen Agent Trust Hub on Jul 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill interacts with the local file system by reading project metadata files (AGENTS.md, README, configuration files, and handoff.md) to gather context and writing generated output (RDQ specification cards) to a local directory named 'rdq/'.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests data from local project files and user attachments to populate its analysis quadrants. If these files contain malicious instructions, they could influence the agent's behavior during the requirement elicitation process.
- Ingestion points: Local files including
AGENTS.md,README,handoff.md, and various project configuration files, as well as user-uploaded attachments. - Boundary markers: The skill uses a structured 'Four Quadrant' framework to organize information, but does not implement explicit delimiters or 'ignore' instructions to isolate untrusted data from the system prompt.
- Capability inventory: The skill possesses the ability to read arbitrary project files, create and modify Markdown files on the local disk, and transfer data to downstream AI skills.
- Sanitization: There is no evidence of input validation or sanitization of the content extracted from local files before it is processed by the agent.
Audit Metadata