dt-rpi-integration

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection Consideration: The skill operates by ingesting and processing Design Thinking artifacts, such as synthesis themes and stakeholder maps, which are produced during earlier stages of a project. This data flow represents a surface where the agent processes external content.
  • Ingestion Points: The agent is instructed to read files located within the .copilot-tracking/dt/{project-slug}/ directory as part of the readiness assessment and artifact compilation steps, as described in subagent-handoff.md and rpi-handoff-contract.md.
  • Capability Inventory: The skill involves reading existing documentation, compiling new handoff artifacts, and validating data across different stages of the Research, Planning, and Implementation cycle.
  • Sanitization Measures: The skill explicitly incorporates "Content sanitization rules" during the validation of generated artifacts to mitigate potential risks associated with external data (specified in subagent-handoff.md).
  • Boundary Markers: While the instructions do not use specific text delimiters for all inputs, the process is structured through multi-step subagent validation to ensure data consistency and integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 06:27 PM
Security Audit — agent-trust-hub — dt-rpi-integration