Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent Markdown/Obsidian knowledge-base helper with disclosed local file maintenance and bounded optional web/media use.
Install this if you want an agent to maintain a long-lived Markdown/Obsidian wiki. Expect it to read and edit files in the target Knowledge vault, maintain indexes and logs, and optionally use web sources or localize images when the task calls for it. Keep private vaults scoped carefully and confirm before bulk renames, migrations, or attachment downloads.
Referenced artifact was not completely inspected
y_wiki.py` as `scripts/verify_wiki.py` inside the skill package, together with `SKILL.md`; the root-level wrapper is only a convenience entry point for this vau
Without declared permissions the skill's intent is opaque and cannot be validated.
The instructions require English as the default for directory and file names, including guidance to keep specific English filenames and prefer readable English names whenever possible. This is a natural-language locale policy constraint that can override user language preferences without an explicit opt-in, which matches the policy-violation criteria.
The skill authorizes external search when certain conditions are met, which expands the agent’s trust boundary from the local knowledge base to the open web. That can introduce prompt-injection content, untrusted facts, and unintended data exposure if queries derived from private wiki contents are sent to external services.
The manifest describes organizing and auditing a Markdown knowledge wiki, but these lines add a capability to download external images and rewrite references into localized attachments. That is a network-fetching and asset-ingestion capability not clearly implied by the skill's stated purpose of maintaining textual knowledge pages.
No suspicious patterns detected.