obsidian-similar-notes
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is designed to perform semantic similarity searches within a local Obsidian vault. Technical analysis of the provided Python script confirms it is limited to mathematical operations (cosine similarity) on local vector data and path resolution. No network activity, credential access, or privilege escalation patterns were identified.
- [INDIRECT_PROMPT_INJECTION]: The skill processes file paths and headings from the local filesystem, which serves as an ingestion point for external data.
- Ingestion points:
scripts/find_similar.pyreads note paths and heading strings from the.smart-env/multi/directory within the Obsidian vault. - Boundary markers: None; the script assumes the integrity of the data provided by the Smart Connections plugin.
- Capability inventory: The script's operations are limited to reading local embedding files and printing results to the console. It contains no capabilities for network communication, subprocess execution, or dynamic code evaluation.
- Sanitization: File paths and heading names are displayed as retrieved from the embedding store without additional sanitization or escaping.
Audit Metadata