custom-architecture
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed to analyze a repository's actual file structure and configuration to generate or refresh architecture documentation in
FRAMEWORK.local.md. It utilizes local file read and write operations that are consistent with its stated purpose.\n- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests untrusted data from the repository (e.g., source code, manifests, and existing documentation) to summarize the architecture. However, the impact is low as it only affects documentation output.\n - Ingestion points:
FRAMEWORK.md,FRAMEWORK.local.md, package manifests, and source directories.\n - Boundary markers: Absent for ingested data; uses
<!-- aircury-custom-architecture -->markers for its own generated output.\n - Capability inventory: Read and write access to the local file system.\n
- Sanitization: No explicit instructions are provided to the agent to ignore or sanitize embedded instructions within the analyzed files.
Audit Metadata