improve-codebase-architecture
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from the project codebase (e.g., CONTEXT.md, ADR files, and source code) to inform its recommendations. An attacker who can commit files to the repository could potentially influence the agent's architectural advice.
- Ingestion points: Project documentation files (CONTEXT.md, CONTEXT-MAP.md, docs/adr/) and source files accessed via the Explore sub-agent.
- Boundary markers: Absent. The instructions do not provide directives to ignore potentially malicious instructions embedded in the analyzed codebase.
- Capability inventory: The skill has the capability to write to the file system (updating documentation) and spawn sub-agents for exploration and design tasks.
- Sanitization: Absent. Content from analyzed files is incorporated into the agent's context and sub-agent briefs without explicit validation.
Audit Metadata