weolopez

Evaluate workflow run quality by analyzing artifacts and scratch outputs, identify gaps and root causes, and write a structured assessment with actionable recommendations.

Write an Architecture Decision Record (ADR) for a technical decision

Diagnose a bug, identify root cause, implement a targeted fix, and verify it

Analyze requirements and produce a detailed implementation plan before writing code

Create a GitHub branch, commit changes to it, and open a pull request

Explore a newly-loaded workspace, map its structure, and write project notes to memory

Restructure code without changing external behavior — extract, rename, split, move

Verify that a previous step's claimed outputs actually exist and are correct — emits PASS, FAIL, or BLOCKED

Manage scoped Claude Code Remote Control instances ("agents"). Each agent is a named, folder-scoped Claude that runs as its own systemd service (e.g. "lucas" works only inside /root/weolopez.github.io/lucas). TRIGGER on "start/launch the <name> agent", "stop/restart the <name> agent", "create/delete an agent", or "list agents". "Start the <name> agent" bootstraps the site via the new-site skill if its folder is missing, then starts the instance.

Create a new *.weolopez.com subdomain — generates the page, wires up static-server.ts, creates the NPM proxy host, and pushes to GitHub. No DNS update needed (wildcard already set).

A test skill for workspace integration testing