design-handoff-reader

Pass

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection attack surface because it ingests and processes untrusted data from external design bundles to generate instructions for other agents.\n
  • Ingestion points: The skill accepts input from local ZIP archives, directories, and external URLs as specified in the '입력' section of SKILL.md.\n
  • Boundary markers: Absent; the instructions do not provide delimiters or specific guidelines for the agent to treat content within the bundles as untrusted data or to ignore embedded commands.\n
  • Capability inventory: The agent performs file globbing, reads Markdown/JSON files (README.md, design-tokens.json, chat-history.md), and generates a production-ready instruction string for the Claude Code agent.\n
  • Sanitization: Absent; no mention of sanitizing, escaping, or validating the content extracted from the handoff bundles before processing it.\n- [EXTERNAL_DOWNLOADS]: The skill is designed to fetch and analyze content from external URLs provided by the user during the handoff process, which is an external data ingestion mechanism.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 19, 2026, 01:42 PM
Security Audit — agent-trust-hub — design-handoff-reader