Security audit
Security checks for vulnerabilities and agentic risk
The skill appears to automate a real AI news digest workflow, but it also schedules recurring runs, reads a local API key, and automatically publishes GitHub changes without enough scoping or user control.
Review this carefully before installing. It is suitable only if you intentionally want a scheduled local automation that reads your MiniMax key, edits the named Jekyll repository, and pushes posts to GitHub without a manual approval step. Prefer adding explicit secret handling, a dry-run mode, confirmation before push, and clear launchd install/remove instructions.
The documented setup requires reading a local secret file and interacting with a local repository path, yet the declared permissions only mention HTTP requests and file writes. This under-declaration is dangerous because it obscures the skill's true access needs, potentially enabling local data access and repository manipulation beyond what a reviewer or runtime policy would expect.
The manifest description presents the skill as a simple news digest, but the body documents materially more sensitive behavior: creating files, auto-committing, and pushing to GitHub. This mismatch can mislead users or orchestration systems about the true side effects of running the skill, reducing informed consent and increasing the chance of unintended repository modification.
The skill performs outbound network activity and uses an external API key, but this is not surfaced as a clear warning in the primary description. Inadequate disclosure of network egress and secret handling can lead to unanticipated data sharing with third parties and poor credential governance.
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.
- Fetches recent HN stories from the past 3 days
- Filters for AI/ML keywords
- Uses MiniMax to pick the 3 most significant and write brief contextual commentary
- Creates a Jekyll post in `_posts/YYYY-MM-DD-midweek-ai-digest.md`
- Auto-commits and pushes to GitHub
The skill description notes file creation and GitHub pushes in the body, but it does not give an explicit warning that running the skill modifies local content and publishes changes remotely. Without prominent disclosure, users may trigger irreversible or externally visible actions they did not expect.
The skill requires a local secret file containing a MiniMax API key, but that sensitive dependency is not reflected in the declared permissions or manifest scope. Hidden secret access increases the risk of unauthorized credential use, accidental exposure, or execution in environments that did not intend to grant local secret-file access.