agent-init-deep

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs legitimate project maintenance tasks using authorized tools (Bash, Read, Write, Edit). No evidence of prompt injection, credential harvesting, or network exfiltration was found.
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests user-controlled data from an existing CLAUDE.md file to perform refactoring. 1. Ingestion points: Reads existing CLAUDE.md during the migration path. 2. Boundary markers: No explicit delimiters or 'ignore' instructions are used when processing the file content. 3. Capability inventory: The skill has file-write (Write/Edit) and shell (Bash) capabilities. 4. Sanitization: No explicit sanitization or validation of the input file content is performed. This vulnerability surface is inherent to the skill's primary function of processing and refactoring documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 06:26 PM