Security audit
Security checks for vulnerabilities and agentic risk
This skill openly creates and shares skills to Slack, but it automatically posts metadata and file links without a clear per-use confirmation or recipient review step.
Install only if you want skill creation to include team Slack announcements. Before use, ensure the Slack workspace and target channel are appropriate, and avoid using it for skills whose names, descriptions, paths, or links reveal sensitive internal information unless you add a manual review step.
The trigger condition is broad enough to activate the skill whenever a user mentions creating or sharing a skill, without clearly requiring informed consent for the Slack-posting behavior. In this context, that ambiguity increases the chance the agent will invoke a workflow that packages artifacts and broadcasts metadata externally when the user only intended local skill creation.
The skill describes automatic Slack sharing but does not prominently warn users that skill metadata and links to skill files may be posted to team channels. This is dangerous because created skills can contain sensitive project names, internal descriptions, or file locations, leading to unintended information disclosure across a broader audience.
The skill is designed to autonomously send information to designated Slack channels, which is an external side effect performed without an explicit per-use approval gate in the specification. Autonomous outbound communication is risky here because it can disclose internal metadata and links, amplify accidental sharing, and remove a human check before publication.
- Runs validation automatically before packaging
### 4. Slack Integration via Rube
- Automatically sends created skill information to designated Slack channels
- Shares skill metadata (name, description, link)
- Posts skill summary for team discovery
- Provides direct links to skill files
No suspicious patterns detected.