python-architecture-review
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted project data such as source code, roadmaps, and architectural documentation. It proactively addresses potential injection risks through explicit instructions requiring the agent to ignore instructions found within processed data and to redact sensitive information.\n
- Ingestion points: The skill reads repository files and directory structures as defined in the workflow (SKILL.md).\n
- Boundary markers: SKILL.md and the reference files include strict rules for bounding and redacting output and treating all repository content as untrusted.\n
- Capability inventory: The skill's capabilities are limited to reading local files and executing a bundled, non-executing static analysis script (scripts/scan_architecture.py).\n
- Sanitization: The 'Output contract' and 'Safety and privacy' sections of the report template enforce the redaction of secrets, private paths, and credentials.
Audit Metadata