swiftlys2-toolkit
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes and audits C# plugin code and external documentation, which constitutes an attack surface where untrusted data enters the agent context.\n
- Ingestion points: User-supplied source code for plugins and external full-text documentation files (
llms-full.txt).\n - Boundary markers: The skill explicitly instructs the agent to request user permission before reading large external files, acting as a manual verification gate.\n
- Capability inventory: The skill provides architectural planning and code analysis; it does not request tools for shell execution, file system modification, or non-whitelisted network operations.\n
- Sanitization: No automated sanitization is specified for user code input, relying instead on manual review and agent safety boundaries.\n- [EXTERNAL_DOWNLOADS]: The skill references official resources and community wikis necessary for framework development.\n
- Fetches technical documentation from the official
swiftlys2.netproject website.\n - References public GitHub repositories for SwiftlyS2 and related community projects (
swiftly-solution/swiftlys2,himenekocn/sw2-mdwiki).
Audit Metadata