Security audit
Security checks for vulnerabilities and agentic risk
This skill creates a local markdown wiki and its file reading, writing, and logging behavior is disclosed and aligned with that purpose.
Install only if you want an agent to maintain a persistent local markdown knowledge base. Avoid placing sensitive documents in llm-wiki/raw/ unless you are comfortable with summaries, derived pages, and query logs being stored in the workspace; ask the agent to identify exact files and preview planned wiki changes before ingestion when working in important repositories.
The skill states that query interactions are logged, but it does not clearly warn users that ordinary questions may be stored persistently. This can lead to accidental retention of sensitive prompts, internal project details, or personal data in log files that may later be read, synced, or exposed.
The invocation language for processing new sources is broad enough that an agent may ingest and act on any newly added file without clear scoping, confirmation, or trust boundaries. In a skill that reads arbitrary documents and then writes multiple persistent files, underspecified triggers increase the risk of prompt injection from source content and unintended modification of the knowledge base.
The skill says it automatically creates the folder structure and base files on first use, but it does not provide a clear up-front warning that using the skill will modify the workspace. Silent or unexpected file creation can violate user expectations, interfere with repositories, and create persistence that later workflows may trust or consume.
No suspicious patterns detected.
