improve-codebase-architecture
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the codebase and documentation files.
- Ingestion points: The skill reads
CONTEXT.md,CONTEXT-MAP.md, and files withindocs/adr/, as well as exploring the codebase using an exploration sub-agent (referenced inSKILL.md). - Boundary markers: There are no explicit instructions or delimiters to isolate content from these external files to prevent them from overriding agent behavior.
- Capability inventory: The agent has the capability to explore the file system and modify documentation files (
CONTEXT.md, ADRs) based on its analysis. - Sanitization: The skill does not describe any sanitization or validation of the content read from the codebase or documentation files.
Audit Metadata