spring-cleaning
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for reading and modifying a codebase, which inherently involves processing untrusted data that could contain malicious instructions.
- Ingestion points: The instructions require the agent to analyze the full codebase, including source files, manifests, and configuration files as described in SKILL.md.
- Boundary markers: Absent. The skill lacks explicit instructions or delimiters to disregard embedded commands found within the codebase being processed.
- Capability inventory: The agent is tasked with deleting files (Phase 1) and restructuring existing code (Phase 2) as outlined in SKILL.md, requiring file system write and delete permissions.
- Sanitization: Absent. There are no sanitization or validation steps defined to prevent the agent from being influenced by instructions hidden in code comments (e.g., TODOs, FIXMEs) or string literals.
Audit Metadata