obsidian-sync
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes agent memory logs which may contain untrusted content from external interactions.\n
- Ingestion points: Memory files are read in SKILL.md and processed by the sync.py utility script.\n
- Boundary markers: The skill does not use specific delimiters or warnings to ignore instructions embedded within memory files.\n
- Capability inventory: The skill utilizes shutil.copy2 for file duplication and allows the agent to write to the user-specified vault path.\n
- Sanitization: The content of the memory files is not sanitized or filtered before being copied to the Obsidian vault.\n- [SAFE]: All file operations are local and originate from the agent's own data or a user-provided path. The code uses yaml.safe_load() for configuration and does not perform any network requests or privilege escalation.
Audit Metadata