solo-sdlc

Audit the .claude/ setup for conflicts, stale instructions, drift from current Claude Code spec.

Watch open PRs, automatically respond to code review comments, keep branches rebased. Use with /loop for continuous PR maintenance.

Commit current changes, push to a feature branch, open a PR with auto-generated description.

Implement frontend pixel-equivalent to a provided mockup. Extracts tokens, builds components in isolation, verifies side-by-side.

Investigate a data source ONCE and save reference to docs/data-sources/. Reuses existing reference if fresh.

MUST BE USED for Django implementation - models, views, DRF, admin, signals, migrations. Use when the project is a Django app or Django REST Framework API. Knows the idiomatic Django patterns and the common pitfalls.

MUST BE USED for Go backend implementation. Builds Go services, libraries, CLIs following idiomatic Go and the project's CLAUDE.md rules. Use when implementing in Go - APIs, workers, CLIs, gRPC services.

MUST BE USED for Python backend implementation (FastAPI, Flask, plain Python services). Use when implementing Python services, APIs, workers, CLIs that don't require Django specifically. For Django projects, prefer backend-django.

MUST BE USED after verification passes and before commit/PR. Multi-perspective review: correctness, security, performance, maintainability. Severity-tagged findings (Critical / High / Medium / Low). Stack-aware (Go, Python, Django, React, Vue).

MUST BE USED before writing any code that touches a new data source (database, API, queue). Investigates the source ONCE and produces a permanent reference at docs/data-sources/. PROACTIVELY check for existing reference before re-investigating. Solves the "Claude keeps re-asking about the database" problem.

Write Architecture Decision Records (ADRs) for any meaningful technical decision. Use this skill whenever an architectural choice is made (database, framework, protocol, library, pattern). Triggers when an architecture decision is being made, when alternatives are being weighed, or when the user asks "why did we use X". Outputs a numbered ADR file under docs/architecture/adr/.

Design the system architecture from an approved PRD. Use this skill in Phase 3 of every feature, AFTER PRD is approved and BEFORE writing any code. Triggers on requests for system design, architecture, design doc, "how to build this", tech choices, or component design. Produces a design doc with C4 diagrams, ADRs, and a solutioning gate check.

Apply LLM coding discipline to every edit - minimum code, surgical changes, surface assumptions, goal-driven execution. Use this skill ALWAYS when writing or editing code, especially when modifying existing files. Triggers on any code-writing task, refactor, bug fix, or feature implementation. Inspired by Karpathy's observations on common LLM coding pitfalls. Prevents overcomplication, scope creep, silent assumptions, and weak success criteria.

Multi-perspective code review (correctness, security, performance, maintainability) on implemented changes. Use this skill AFTER verification passes and BEFORE the work is shipped. Triggers when implementation is complete, before merging, or when user asks for review. Outputs severity-tagged findings with concrete fixes, not generic advice.

Generate concrete, actionable improvement recommendations after every phase, every implementation, and every review. Use this skill at the END of every Claude Code task. Triggers when a phase completes, after implementation is done, after a code review, or when the user asks "what could be better". The user explicitly values critical analysis over validation. Outputs structured recommendations the user can act on or ignore.