claude-migrate-memory-to-doc

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs several file system operations using shell commands including mkdir, ln, rm, sed, grep, find, and cp to manage the migration and configuration updates.
  • [COMMAND_EXECUTION]: It executes the codex exec command to verify that the migrated instructions are correctly injected into the tool's context.
  • [DATA_EXFILTRATION]: The skill accesses sensitive local directories such as ~/.claude and ~/.codex, which contain personal memory, project configuration, and session history.
  • [PROMPT_INJECTION]: Ingestion points: Accesses and reads files located in ~/.claude/projects/*/memory/ to migrate their content. Boundary markers: No explicit delimiters or instruction-ignore guards are used when reading or processing these memory files. Capability inventory: The skill can execute shell commands and orchestrate multiple subagents to review the migrated content. Sanitization: There is no evidence of sanitization or filtering applied to the content of the memory files during the migration process.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 05:05 AM
Security Audit — agent-trust-hub — claude-migrate-memory-to-doc