By uta2000
Automate quality gates and documentation tracking during feature development by running static analysis on edits, blocking unsafe writes, and maintaining a commit ledger with impact warnings.
Cleans up worktrees, local branches, remote branches, and handoff files for merged or closed PRs. Invoked automatically by merge-prs after a successful merge, and opportunistically by start: at session kickoff. Safe to run manually at any time — all actions are idempotent.
Consult the codex MCP server as a second AI opinion via a three-phase Claude-orchestrated flow. Use for proactive reviews (design, plan, code) or reactive stuck recovery. Requires codex.enabled=true in .feature-flow.yml and a configured codex MCP server.
This skill should be used when the user asks to "create an issue", "create a GitHub issue", "open an issue", "write up an issue", "file an issue", or after a design has been verified and needs to be captured as a trackable GitHub issue for implementation. Also handles updating an existing issue when an issue number is provided as context.
This skill should be used when the user asks to "write a design doc", "create a design document", "document the design", "write up the design", "write a spec", "spec this out", or when brainstorming is complete and decisions need to be captured before implementation planning.
This skill should be used when the user asks to "verify the design", "check the design against the codebase", "validate the design doc", "check for blockers", "will this design work", or after writing a design document to catch schema conflicts, type mismatches, and pipeline assumptions before implementation, "check assumptions", "what am I assuming", "verify assumptions", "surface assumptions".
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that enforces a full feature development lifecycle — from idea to merged PR — so you don't discover schema mismatches, deprecated APIs, or broken assumptions halfway through coding. You say "start: add user notifications" and it auto-detects your tech stack, queries live documentation for current patterns, walks you through brainstorming and design, then verifies that design against your actual codebase (columns, types, routes, constraints) before a single line of implementation is written. After implementation (done via TDD in an isolated git worktree), it runs a multi-agent code review pipeline with up to 7 specialized reviewers, auto-fixes findings, generates a CHANGELOG entry, and mechanically verifies every acceptance criterion before opening a PR.
The upfront design adds ~20-30 minutes but typically saves 2-4 hours of mid-implementation debugging per feature. It works with any tech stack — Next.js, Supabase, Django, Rails, Flutter, whatever — and gets smarter over time by writing discovered gotchas (like "PostgREST silently caps queries at 1000 rows") back into your project config so every future feature is checked against past lessons.
Five features bundled into this release — all addressing gaps in design review and mid-lifecycle judgment that the existing gate chain missed. See CHANGELOG.md for the detailed entries.
| PR | Issue | What it adds |
|---|---|---|
| #243 | #238 | Scope-critique pass — new Step 4.5 in design-verification asks five strategic-shape questions (scope, dependencies, simpler alternatives, observability, config surface) before a design reaches create-issue. Catches oversized scope, phantom dependencies, and unobservable capability bets. |
| #242 | #236 | Advisor tool integration — four soft advisor-call hints at judgment-heavy moments, plus a one-line SessionStart onboarding nudge for Sonnet users missing the beta header. Opt-in; see docs/advisor.md. |
| #241 | #239 | Senior developer panel (Phase 1c) — Major-feature code review gains a judgment-review phase with three personas (Staff Engineer, SRE, Product Engineer). Surfaces wrong abstractions, operability risk, and scope creep. Findings are report-only — you decide whether to address, defer, or reject. |
| #240 | #235 | Codex consultation (Phase 1+2, opt-in) — new consult-codex skill for second-opinion AI reviews via the existing codex MCP server. Ships review-design proactive mode wired into design-document; later modes (review-plan, review-code, stuck) deferred. Disabled by default; enable via codex.enabled: true. |
| #237 | #234 | Quick-path triage — start: now routes trivial changes (prose edits, non-log string literals, comments) to a bare implement-and-commit flow, skipping brainstorm / design / verify / plan / handoff. Gated by five code-aware confirmation gates (clean tree, concrete target, file/line budgets, exported-declaration overlap, lexical-region rule). |
gh on PATH, authenticated via gh auth login) — used for issue creation, PR checks, review bot handling, and issue commenting/closingRecommended (for full code review coverage):
# Install required plugins first
claude plugins add superpowers
claude plugins add context7
# Install feature-flow
claude plugins add feature-flow
npx claudepluginhub uta2000/feature-flow --plugin feature-flowConsult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, plus codex, antigravity, and grok CLIs when installed) to get diverse perspectives on coding problems
Production-grade engineering skills for AI coding agents — covering the full software development lifecycle from spec to ship.
Evidence-gated AI coding workflow: scan → analyze → plan → TDD → execute → fix → verify → review, powered by Codebase Memory MCP >= 0.9.0 with optional Serena LSP intelligence. Includes blast-radius planning, test/cycle gates, independent review, and Windows Git Bash hook auto-resolution.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.
v9.52.0 - Reliability wave: tangle contextual review correction loop with hard round ceiling, progress-supervised review rounds (per-agent stall watch, descendant-tree kills), council diversity and agy pin fixes, marketplace generator source-of-truth fix, provider troubleshooting runbook and cost-expectations docs. Run /octo:setup.
Lazy senior dev mode. Forces the simplest, shortest solution that actually works: YAGNI, stdlib first, no unrequested abstractions.