code-refactoring-context-restore
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the way it handles external context retrieval.
- Ingestion points: The skill ingests untrusted data from
context_source(specifically identified as vector databases and file systems) as described in the Core Requirements section. - Boundary markers: There are no instructions to use delimiters or explicit 'ignore embedded instructions' warnings when incorporating retrieved context into the agent's active memory.
- Capability inventory: The skill includes functions like
load_componentandVectorDatabase.searchin SKILL.md which perform file system and database read operations. - Sanitization: The instructions do not specify any validation, escaping, or filtering of the content retrieved from the
context_sourcebefore it is rehydrated into the prompt context.
Audit Metadata