neptunehub
- 3 skills
- 0 followers
- 18 hours ago last updated
- ▌ Sec Review · neptunehubSecurity review of this codebase - reasons about the code like a security researcher, tracing untrusted input from entry point to sink rather than pattern-matching. Scans the whole tree at HEAD for hardcoded secrets (anything beyond the empty defaults config.py declares), then checks what the current change introduces against the head of remote main. Uses the OWASP Top Ten as the primary risk source, reports severity plus confidence in a markdown table, and proposes fixes without ever applying them. Use when asked whether the code is secure, to audit for vulnerabilities, to check for injection, auth, secrets, or dependency risk, or to review a change for security regressions.
- ▌ Arch Review · neptunehubScan the codebase for architectural friction and return an action plan - a ranked list of deepening opportunities, each scored for regression risk from 1 to 10, ordered by the value it adds. Refactors that turn shallow modules into deep ones, for testability and AI-navigability. Output is a markdown table plus pointed lists, never HTML and never a written report file. Use when the user asks where the codebase is getting hard to change, for refactoring opportunities, or for an architecture review.
- ▌ Code Review · neptunehubFast, project-specific review of a target - the working tree (default), a commit, a range, a branch, or a PR - against the head of remote main. Hunts regressions in callers, task-queue and multi-server contract breaks, build-matrix breaks, 1M-song scale and idle-RAM defects, and breaches of this project's hard rules. Output is ONE table - change requested, priority, risk of regression. Use when the user wants a commit, branch, PR, or work in progress reviewed.