lucaswall
- 15 skills
- 0 followers
- 6 hours ago last updated
- ▌ Roadmap · lucaswallDeep research and discussion of a roadmap feature or new idea. Gathers extensive context from codebase, web, APIs, MCPs, and project history, then presents a concise analysis report for interactive discussion. After discussion, handles the outcome — write a feature spec to the roadmap, move to backlog, create an inline plan, modify, or drop. Use when user says "roadmap", "pull from roadmap", "push to roadmap", "add to roadmap", "analyze this feature", "research this idea", or wants to evaluate a feature.
- ▌ Plan Fix · lucaswall bundleInvestigates bugs AND creates actionable TDD fix plans. Creates Linear issues in Todo state. Use when you know you want to fix something - user reports errors, deployment failures, wrong data, or UI issues. Can be chained from investigate skill. Discovers MCPs from CLAUDE.md for debugging (logs, etc.).
- ▌ Code Audit · lucaswall bundleAudits codebase using an agent team with 3 domain-specialized reviewers (security, reliability, quality). Triages open Sentry issues (creates Linear issues for real bugs, resolves/ignores noise). Creates Linear issues in Backlog state for findings. Use when user says "audit", "find bugs", "check security", "review codebase", or "team audit". Higher token cost, faster and deeper analysis. Falls back to single-agent mode if agent teams unavailable.
- ▌ Deep Review · lucaswall bundleDeep, focused analysis of a single screen or feature. Combines code correctness, security, UX, and performance in one unified Opus pass with cross-domain reasoning. Finds bugs that broad audits miss by tracing full data flows and component interactions. Use when user says "deep review", "deeply analyse", "review this screen", "find all bugs in X", or wants thorough analysis of a specific area. Requires a target area argument.
- ▌ Investigate · lucaswallRead-only investigation that reports findings WITHOUT creating plans or modifying code. Use when user says "investigate", "check why", "look into", "diagnose", or wants to understand a problem before deciding on action. Accesses deployment logs and codebase.
- ▌ Plan Inline · lucaswall bundleCreate TDD implementation plans from direct feature requests. Use when user provides a task description like "add X feature", "create Y function", or "implement Z". Creates Linear issues in Todo state. Faster than plan-backlog for ad-hoc requests that don't need backlog tracking.
- ▌
- ▌ Tools Improve · lucaswall bundleREQUIRED: Load this skill BEFORE creating, modifying, or reviewing any Claude Code skill, subagent, or CLAUDE.md file. Contains critical best practices for .claude/skills/, .claude/agents/, and CLAUDE.md. Use when: creating agents, creating skills, editing SKILL.md files, editing agent .md files, reviewing skill/agent code, reviewing or editing CLAUDE.md, adding instructions to CLAUDE.md, or any work involving Claude Code extensibility.
- ▌ Tools Migrate · lucaswallMigrate skills and agents between this plugin and another project. Use when: syncing workflow tools to a target project, pulling improvements back from a project, comparing skill versions, or adapting tools to a different architecture. Supports bidirectional migration with architecture adaptation.
- ▌ Add To Backlog · lucaswallAdd issues to Linear Backlog from free-form input. Use when user says "add to backlog", "create backlog issues", "track this", or describes tasks/improvements/bugs to add. Interprets user's ideas, investigation findings, or conversation context into well-structured Backlog issues. Can process multiple items at once.
- ▌ Backlog Refine · lucaswallRefine vague Linear Backlog issues into well-specified, actionable items. Use when user says "refine backlog", "refine PROJ-123", "improve backlog items", or "clarify issues". Fetches Backlog issues, analyzes gaps, asks clarifying questions, and updates Linear after user confirms.
- ▌ Plan Implement · lucaswall bundleExecute the pending plan in PLANS.md using an agent team for parallel implementation. Use when user says "implement the plan", "execute the plan", "team implement", or after any plan-* skill creates a plan. Spawns worker agents in isolated git worktrees for full code isolation. Updates Linear issues in real-time. Falls back to single-agent mode if agent teams unavailable.
- ▌ Frontend Review · lucaswall bundleReviews all frontend elements (UI, UX, accessibility, visual design, responsiveness, performance, visual QA from screenshots) using an agent team with 4 domain-specialized reviewers. Creates Linear issues in Backlog state for findings. Use when user says "review frontend", "check UI", "review UX", "audit accessibility", "check responsive", or "review screens". Falls back to single-agent mode if agent teams unavailable.
- ▌ Push To Production · lucaswall bundlePromote main to release with production DB backup and migration handling. Use when user says "push to production", "release", "deploy to production", or "promote to release". Backs up production DB, assesses MIGRATIONS.md, writes migration code if needed, and merges main to release.
- ▌ Plan Review Implementation · lucaswall bundleQA review of completed implementation using an agent team with 3 domain-specialized reviewers (security, reliability, quality). Use after plan-implement finishes, or when user says "review the implementation". Moves Linear issues Review→Merge. Creates new issues in Todo for bugs found. Falls back to single-agent mode if agent teams unavailable.