migration-loop
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the local repository and a configuration file (migration-checklist.json). This creates a surface where malicious instructions embedded in the code or configuration could influence agent behavior.
- Ingestion points: Repository content via Grep, Glob, and Read; configuration via migration-checklist.json (SKILL.md, references/migration-checklist.md).
- Boundary markers: Uses structured JSON for configuration, but lacks explicit prompt delimiters for source code content.
- Capability inventory: Bash (command execution), Write/Edit (file modification), and Agent (sub-agent spawning) as defined in SKILL.md.
- Sanitization: No explicit sanitization of the code being migrated is described in the skill files.
Audit Metadata