Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed local memory and guard wrapper for Vestige, with persistence that matches its stated purpose.
Install this only if you want a separate local Vestige MCP memory store. Avoid saving secrets or confidential content, consider using a per-project --data-dir for containment, and review the upstream npm package before global installation.
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
vmem recall <tag-or-id> # exact handle
vmem save "text" [tag] # smart_ingest, optional tag
vmem walk <failure-memory-id> # causal_walk from a logged write
vmem receipt <memory-id> # receipt get
vmem selftest
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
vmem recall <tag-or-id> # exact handle
vmem save "text" [tag] # smart_ingest, optional tag
vmem walk <failure-memory-id> # causal_walk from a logged write
vmem receipt <memory-id> # receipt get
vmem selftest
In the save command, the script takes arbitrary user content and sends it to the smart_ingest tool via vestige-mcp. While the usage text names the command, there is no explicit warning, confirmation, or comment disclosing that supplied content will be transmitted to another process/service for storage.
This markdown file describes that the skill 'saves that chain to memory' and records repository history, but the user-facing description does not explicitly warn that using these tools persists data locally and could store sensitive project information if misused. Although there is a warning not to save secrets, the broader persistence/privacy impact of writing session and codebase-related data is not clearly disclosed as a caution.
No suspicious patterns detected.