design-sync
Pass
Audited by Gen Agent Trust Hub on Feb 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (LOW): The skill invokes several shell commands including
ls,grep,sed,git, andgo list. These are used to inspect the repository structure and history. While these commands are standard development tools, the use of shell execution (!) is a high-privilege capability that should be monitored.\n- [DATA_EXFILTRATION] (LOW): The skill accesses internal project files such asDESIGN.mdand source code within theinternal/directory. This constitutes data exposure to the LLM. No external network requests were detected, limiting the risk to the agent context.\n- [PROMPT_INJECTION] (LOW): There is a surface for indirect prompt injection (Category 8) because the skill processes content fromDESIGN.md.\n - Ingestion points:
DESIGN.md(read viased,grep, andhead).\n - Boundary markers: Absent; the content is piped directly into analysis instructions without delimiters.\n
- Capability inventory: Local shell execution (
ls,sed,grep,git,go).\n - Sanitization: Absent; the skill does not escape or validate the contents of the documentation before processing.
Audit Metadata