code-refactoring-context-restore
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill describes a framework for 'Context Restoration' that ingests data from external sources like vector databases and file systems. This architecture inherently creates a surface for indirect prompt injection if the ingested content contains adversarial instructions. Evidence Chain: 1. Ingestion points: The
context_sourceparameter andsemantic_context_retrievefunction indicate data entry from external storage. 2. Boundary markers: No delimiters or explicit instructions to ignore embedded commands are present in the provided markdown. 3. Capability inventory: The documentation describes semantic search and context ranking but does not include executable scripts that perform high-risk operations. 4. Sanitization: No evidence of input validation or sanitization of retrieved context is documented. - [Prompt Injection] (SAFE): No instructions designed to bypass agent safety filters or override system-level constraints were detected.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials (API keys, tokens) or unauthorized network exfiltration patterns were identified.
- [Remote Code Execution] (SAFE): No patterns involving the download and execution of remote scripts (e.g., curl|bash) were found.
- [Privilege Escalation] (SAFE): The skill does not request or utilize elevated permissions such as sudo or administrative access.
Audit Metadata