Profile My skills Saved Collections Edit profile Submit a skill

← all publishers

mrveiss

@mrveiss source repo

20 published skills

  1. ▌
    Commit · mrveiss
    Standardized commit workflow with pre-flight checks, auto-format, and retry logic for pre-commit hooks
    0
    installs
  2. ▌
    Process · mrveiss
    How to approach work before writing code — exploring a request before building, planning a multi-step change, debugging a failure methodically, verifying a claim before calling it done, dispatching parallel agents, and finishing a branch. Use at the START of any non-trivial task, when a bug resists the obvious fix, when about to claim something works, or when a plan needs writing. Not for the domain work itself.
    0
    installs
  3. ▌
    Gap Audit · mrveiss
    After completing a batch fix or implementation, audit adjacent files for the same issue and file GitHub discovery issues for each gap found. Use after any batch implementation, bug fix, or remediation campaign.
    0
    installs
  4. ▌
    UI Design · mrveiss
    Design, build, review, or polish any user interface — visual direction, typography, color, layout, spacing, motion, accessibility, responsive behaviour, theming, UX copy, empty and error states, and reusable design systems. Use for landing pages, dashboards, product UI, app shells, components, forms, onboarding, and redesigns; also for making a bland design bolder, a loud one quieter, or auditing an existing interface. Not for backend-only work.
    0
    installs
  5. ▌
    Web Audit · mrveiss
    Full security, SEO, and AI-friendliness audit for any website. Runs multi-pass checks covering HTTP headers, DNS/subdomain recon, port scan, TLS, CORS, email spoofing, exposed admin panels, plugin CVEs, open resolvers, FTP, per-page SEO, performance, and compromise indicators. Outputs a self-contained HTML report.
    0
    installs
  6. ▌
    Review Lenses · mrveiss
    Domain review lenses — architecture, delivery, frontend, documentation, UX, and visual craft. Use for any review task where the question is "is this good?" rather than "does it run?". Pair with the project's own review process docs, which cover mechanics (dispatch, gates, merge).
    0
    installs
  7. ▌
    Memory Cleanup · mrveiss
    End-of-session memory hygiene ritual. Run this at the end of every session to keep MEMORY.md an index and every memory reachable from a topic file.
    0
    installs
  8. ▌
    Canonical Coding · mrveiss
    The canonical-source discipline for ALL code changes in this repository. Use this skill whenever writing, modifying, reviewing, or planning ANY code — frontend, backend, infra, migrations, or docs that describe code. Especially use it when tempted to create a new helper, wrapper, framework, endpoint, config, schema bootstrap, or "v2" of anything that already exists; when a frontend and backend disagree; when a plan or doc disagrees with the code; or when two implementations of the same concept are found. This skill defines which source wins every conflict.
    0
    installs
  9. ▌
    Pr · mrveiss
    Create a pull request with pre-flight branch checks, targeting Dev_new_gui by default
    0
    installs
  10. ▌
    Drain · mrveiss
    Pick and solve the backlog issues that need no decision. Use when asked to work the backlog, drain issues, "solve what you can", run the loop autonomously, or whenever a loop tick has no assigned issue. Selects work; delegates execution to batch-implement.
    0
    installs
  11. ▌
    Review · mrveiss
    Run a PR review cycle — CI diagnosis, a three-angle finder pass, lint-only auto-fix, and the merge decision. Use when reviewing open PRs, working the review queue, or when a heartbeat tick has no assigned issue.
    0
    installs
  12. ▌
    Implement · mrveiss
    End-to-end GitHub issue implementation — umbrella gate, worktree, design, code, verify, PR, CI, and the three-gate closure check. Use when told to implement, fix, or solve a specific issue number, or when picking the next issue off the backlog. Supersedes the former `issue` skill.
    0
    installs
  13. ▌
    Github CLI · mrveiss
    Use when performing any GitHub operation — issues, PRs, comments, labels, reviews, merges, file contents, branch management, or repository queries. Always prefer gh CLI over GitHub MCP tools.
    0
    installs
  14. ▌
    Review Fleet · mrveiss
    Dispatch a 10-angle parallel PR review fleet (finder agents + verifier agents) that posts only confirmed, deduplicated findings to a single PR comment. Use when asked for a high-effort or recall-biased PR review.
    0
    installs
  15. ▌
    Debug Autobot · mrveiss
    Debug any AutoBot failure across the full stack — dispatches parallel investigators per layer (Vue, FastAPI, Redis, ChromaDB, NPU, Browser, AI Stack), then synthesises their reports into a single root cause. Use for any bug whose layer is unknown, any cross-layer symptom, or when a fix in one layer did not hold.
    0
    installs
  16. ▌
    Batch Implement · mrveiss
    Full implement→review→merge→close→discover loop for GitHub issues. Use this whenever the user says "implement all X-labeled issues", "fix all Y bugs", "run a batch on these issues", or gives a list of issue numbers to work on. Do NOT ask for scope clarification — just run the pre-flight and start.
    0
    installs
  17. ▌
    Dead Code Audit · mrveiss
    Systematic codebase audit for unwired code — identify unregistered routers, uninvoked hooks, orphaned components, and file discovery issues. All findings become "wire it in" issues — never deletion issues.
    0
    installs
  18. ▌
    API Wiring Audit · mrveiss
    Audit and enforce frontend/backend API contract wiring in AutoBot-AI (or any FastAPI + SPA monorepo). Use this skill whenever the user mentions unwired features, dead buttons, 404s from the GUI, API contract drift, endpoints that "don't exist", unmounted routers, frontend/backend mismatch, or before marking ANY frontend feature complete. Also use it when implementing or modifying any frontend API call, any FastAPI route, or any router registration — run the audit as a completion gate, not just when explicitly asked.
    0
    installs
  19. ▌
    Session Lifecycle · mrveiss
    Mandatory start-of-session and end-of-session protocol for every Claude Code session in this repository. Use this skill at the BEGINNING of every session (worktree setup, stale-branch inheritance cleanup, rebase onto latest base) and at the END of every session or when the user says wrap up, finish, end session, hand off, or when the mission is complete or cannot proceed. Also use when resuming or taking over a previous session's work, or when the working tree contains another session's leftovers. This skill applies to ALL sessions regardless of task — coding, docs, audits, reviews.
    0
    installs
  20. ▌
    Pre Merge Validate · mrveiss
    Validate code before merging — syntax, imports, call-site impact, tests, types, and linting
    0
    installs