Security audit
Security checks for vulnerabilities and agentic risk
This skill appears intended to save coding-session notes to Obsidian, but its automatic sync rules can persist detailed session content and file paths without a clear per-write confirmation step.
Install only if you are comfortable with session plans, implementation summaries, file paths, and technical details being written into your configured Obsidian vault. Avoid enabling the optional automatic CLAUDE.md workflow unless you want syncs to happen from broad session-completion cues, and review notes for secrets or confidential content before relying on automatic exports.
The skill advertises convenience and zero-config behavior but does not clearly warn that it will write session-derived Markdown and Canvas artifacts into the user's Obsidian vault. That omission can cause users to invoke it without realizing project data, plans, summaries, or file-path details are being persisted outside the immediate coding environment.
The skill instructs the agent to extract plan content from the current conversation when no plan file exists and sync it into external notes. This creates a direct path for copying sensitive session content, including secrets, proprietary design details, or user-provided data, into the Obsidian vault without meaningful minimization or consent checks.
Mandating preservation of original plan content in full encourages wholesale replication of session material into external storage. In security terms, this defeats data minimization and increases the likelihood that credentials, internal architecture details, incident notes, or other sensitive content are persisted and later exposed.
The rules require never trimming original plan details and always reporting specific file paths, normalizing broad retention of session-derived information and operational metadata. In the context of automatic sync to an external vault, this materially raises disclosure risk by persisting implementation details and filesystem structure that may be sensitive or identifying.
The auto-sync trigger guidance uses broad conversational signals such as approval-like phrases and session-completion heuristics, which can cause the skill to run when the user did not intend external synchronization. Because the skill performs filesystem writes to an external vault, ambiguous triggers increase the risk of unintended data export and persistence.
No suspicious patterns detected.
