vindm
UserAI dev infrastructure framework for Claude Code. /dotclaude:bootstrap authors CLAUDE.md + docs/ + .claude/ tuned to your project. Per-domain: design (showpiece), coding, planning, testing, data, ai-workflow.
Categories
Indexed Skills (19)
authoring-skills
How to write a skill or rule that survives refactors instead of rotting — bind to durable invariants and point at canonical sources, never mirror perishable snapshots (file:line cites, step-by-step flow prose, exhaustive rosters, copied code, bare counts). Load whenever you author or edit a skill, a rule, or any reference doc that names code paths, functions, components, or flow stages.
decomposition
Decompose a file that has grown over the size ceiling into focused sibling files — extracting at cognitive seams (hooks, sub-components, pure helpers, types), in the right order, behind a user-approved plan, with post-extraction verification. Invoke when a file-size check warns or a file has become too large to navigate. Refuses on auto-generated, snapshot, and fixture files.
element-reuse
Before reusing an existing string, component, or copy pattern on a NEW surface — locate its current usage with grep, classify both the existing and the proposed surface by type, and apply the verdict matrix. Catches first-touch copy leaking onto a daily-driver surface, a bug invisible to lint, code review, and per-surface visual review. Load at design time and audit time on any borrowed user-visible element.
handoff
Consciously capture in-flight state before context loss — a /clear, an auto-compaction, or the end of a long session. Routes durable facts to memory, plan progress to the plan doc, orphan WIP to an ephemeral handoff doc, and WIP-commits rather than stashing. Invoke before clearing context, when context is filling toward auto-compaction, or at the end of a long working session with state still in flight.
iterative-polish-autoloop
Drive a multi-screen flow from raw to award-quality through a bounded iterative loop — reset to a clean fixture, capture every state, grade across three scrutiny layers (reviewer agent + composition scan + backend-truth probe), apply ONE highest-ROI fix, commit atomically, re-run — until award-quality OR a hard iteration cap. Invoke for iterative UI/UX polish on a flow with an auto-capture harness.
journey-mapping
Build the prior-surface inventory before designing OR auditing any screen — walk every surface the user touched before reaching this one, classify each by type, and apply the forbidden-pattern matrix so first-touch copy/voice/chrome never leaks onto a daily-driver surface. Load before any new design spec or any single-screen / flow audit.
knowledge-layers
The authority order across a project's three knowledge layers — guidance (.claude/), code (truth), and docs/ (reflection + transient intent) — with code as the tiebreak when a doc and the code disagree. Load when reading or trusting a doc to make a code decision, when authoring rules or skills that reference docs, or whenever a load-bearing fact comes from a description rather than from the code itself.
memory-system
Use the project's cross-conversation memory directory well — typed entries (user / feedback / project / reference), each type with its own lifecycle and promotion path, kept as facts not essays, navigated through a curated index. Load when saving a durable fact, reading prior-session context, or deciding whether something belongs in memory at all.
migration-create
Author a database schema migration safely — one atomic change per migration, classified by risk with an explicit confirmation gate on destructive changes, access policies composed in the same migration as any new table, generated artifacts regenerated after, and a smoke test before declaring done. Discovers the project's own migration tool, conventions, and regen command at runtime. Use whenever you need to change the schema.
operating-discipline
Core operating discipline for any non-trivial task — understand before building, reason to the right solution, execute completely and verified, stay lean, and avoid the parallel-path and subagent-dispatch traps. Load this on any substantive coding, design, planning, debugging, refactor, or review work (skip only for pure mechanics like a typo or an explicit one-liner). The universal "how to work" base, consumed as-is.
persona-testing
Run three orthogonal outside-eyes tests on every user-visible string — day-30 (frequency-jaded), partner (voice-register), and stranger (already-knows-me) — to catch voice violations that pass a deny-list but still feel wrong. Load at design time on every proposed string and at audit time on every visible string on a captured surface.
saturday-ritual
Run an OPTIONAL bounded weekly drift-detection cadence — batch the findings hooks can't prevent (stale skills, stale docs, dead rules) into one decision sheet the user marks fix-now / defer / need-info / won't-fix, tracked in a single canonical open-findings registry. Small, young, or solo projects skip this entirely. Use only when a long-lived project has accumulated review-worthy drift.
ai-workflow
Set up LLM workflow discipline for projects that use AI / LLM calls in production or eval suites. Authors an eval-cost-watcher agent that projects token cost BEFORE regression evals run, plus an AI-workflow-discipline rule covering mock-mode placement, fixture freshness, and multi-stage cost accumulation. Optionally a small eval-regression skill stub. Invoke /dotclaude:ai-workflow only when the project actually has AI workflows.
coding
Set up coding hygiene + voice discipline for a project. Authors a tailored kit of code-review agent, decomposition skill, file-size + voice rules, and edit-time hooks — derived from the project's stack, its actual file-size distribution, the bug classes visible in git history, and any project-specific voice. Invoke /dotclaude:coding in any project root with source code.
data
Set up database / persistent-state discipline for a project with a DB. Authors a data-auditor agent (5 categories of integrity check), a migration-create skill, query-discipline rule (CLI > MCP DB tool for reads), and edit-time regen hook — all derived from the project's actual schema, migration framework, and access-control model. Optionally an RLS-security-reviewer if Postgres + row-level security is in use. Invoke /dotclaude:data in any project with persistent state.
design
Set up design / IA / UX / a11y / visual-quality discipline for a project. Authors a tailored kit of design audit agents, IA skills, and design-token rules — derived from the project's actual UI surfaces, the user's named design benchmarks, and past design failure modes. Invoke /dotclaude:design in any project root that has a human-facing surface.
init
Lighter alternative to /dotclaude:bootstrap when you want Layer 6 domain kits only, skipping upstream layers. Orchestrates the relevant subset of domain skills (design / coding / planning / testing / data / ai-workflow) based on project shape — no identity / architecture / process / quality-bar / knowledge-graph / maintenance authoring. Use /dotclaude:init for incremental Layer 6 setup on partial-brownfield projects, or when the user already has CLAUDE.md authored by hand.
planning
Set up pre-implementation discipline + audit routing for a project with regular multi-module changes. Authors a pre-flight agent that maps integration points, parallel paths, and cross-boundary surfaces — plus an audit-routing rule that orders which agent fires for which question. Invoke /dotclaude:planning in any project where "we shipped feature X and broke unrelated thing Y" has happened more than once.
testing
Set up test architecture + coverage discipline for a project. Authors a tests-architect agent (4 modes — audit / design / implement / maintain) tuned to the project's stack, test framework, risk model, and existing patterns. Plus a small "test discipline" rule when helpful. Invoke /dotclaude:testing in any project where tests exist OR should exist.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.