dot-ai-query-dot-ai

Warn

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: MEDIUMDATA_EXFILTRATION
Full Analysis
  • [DATA_EXFILTRATION]: The skill explicitly instructs the agent to traverse the file system to sibling directories (../) to access multiple other projects (dot-ai, dot-ai-ui, dot-ai-controller, etc.). This expands the agent's access scope beyond the immediate project environment.
  • [DATA_EXFILTRATION]: The skill identifies dot-ai-infra as containing "Production infrastructure" and permits the agent to query it upon user request. Accessing infrastructure code significantly increases the risk of exposing sensitive deployment configurations or secrets.
  • [DATA_EXFILTRATION]: The instructions encourage tracing code paths from "definition → implementation → exposure," which involves deep recursive reading of files across multiple projects, increasing the likelihood of encountering and potentially leaking sensitive internal implementation details.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it is designed to ingest and process untrusted data from external codebases.
  • Ingestion points: Reads files from sibling project directories (../dot-ai, ../dot-ai-ui, etc.) using Grep, Read, and Glob tools.
  • Boundary markers: Absent. No instructions are provided to the agent to ignore or delimit instructions found within the code comments or documentation of the queried projects.
  • Capability inventory: The agent has full access to file system tools (Read, Grep, Glob) and has no restrictions on tool usage in the YAML frontmatter.
  • Sanitization: Absent. There is no validation or filtering mentioned for the content retrieved from the external projects.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 24, 2026, 01:25 AM
Security Audit — agent-trust-hub — dot-ai-query-dot-ai