By shiyas1331
A Claude Code plugin with debugging, code analysis, and developer productivity commands.
Read reviewer comments on a PR, classify them, draft fixes for change-requests and replies for questions, commit and post — with author approval at every step
Address PR feedback and automatically mark threads resolved after fixes land — skips the per-thread "mark resolved?" prompt. Skips the picker.
Address PR review feedback — walk through each comment with confirmation, commit fixes, post replies. Default mode when invoked with a bare PR identifier.
Show the plan only — classify comments and propose actions, but DON'T apply changes or post anything. Skips the picker.
Show the verbose flag reference for /devkit:address-pr.
Analyzes codebase implementation details. Call when you need to understand how specific components work, trace data flow, or map execution paths. The more detailed your prompt, the better the analysis.
Locates files, directories, and components relevant to a feature or task. Call with a human language prompt describing what you're looking for — a "Super Grep/Glob/LS tool" for when you need to search across multiple patterns.
Finds similar implementations, usage examples, or existing patterns that can be modeled after. Like codebase-locator but also extracts concrete code examples and shows how patterns are used.
Checks a code diff against a repo's documented conventions (CLAUDE.md, .claude/codebase/*.md, CONTRIBUTING.md, STYLE.md). Surfaces matches and deviations with file:line references and severity. Reusable across pr-review, address-pr, and pre-PR review tools.
Writes one unit test file for one source file, runs it, and retries failures. Per-file scope. Always operates within a platform adapter loaded by /devkit:cover. Returns structured JSON describing what landed.
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 with debugging, code analysis, and developer productivity commands.
Run these commands in your terminal:
claude plugin marketplace add shiyas1331/devkit
claude plugin install devkit@shiyas-devkit
Restart your Claude Code session. The plugin is now available.
If you want to make changes and test locally:
git clone https://github.com/shiyas1331/devkit.git
cd devkit
claude --plugin-dir .
When a new version is released:
claude plugin marketplace update shiyas-devkit
claude plugin update devkit@shiyas-devkit
Changes take effect in your next Claude Code session.
claude plugin uninstall devkit@shiyas-devkit
claude plugin marketplace remove shiyas-devkit
/devkit:why — Why does this code exist?Replaces the daily 5-minute git blame + GitHub-clicking dance with a 10-second answer. Tells you who added the code, when, in which PR, with what reasoning. Pulls blame, originating commit, PR description, linked JIRA ticket, and review-thread highlights into one grounded explanation.
Walks line history. The most recent commit on a line is often a typo fix or formatter run — not the answer to "why." This command walks back through history (skipping trivial commits like typo/format/lint/rename/bot edits) to find the substantive originator. The walk count is shown so you know how far it traced.
Usage:
/devkit:why packages/editors/src/api/apiClient.ts:188
/devkit:why apiClient.ts # entire file's origin
/devkit:why apiClient.ts:120-150 # block of lines
/devkit:why apiClient.ts:188 --depth=thorough # full drilldown
/devkit:why apiClient.ts:188 --max-walk=10 # walk further back
/devkit:why apiClient.ts:188 --json # machine-readable output
Detects special cases:
git log --follow)merge_method)Mechanical confidence labels (no judgment-call variance):
high — PR body or commit message explicitly explains the why (because, to fix, to address, etc.). OR ticket description >100 chars.medium — Linked ticket has meaningful title. OR review thread has ≥2 rationale comments. OR commit body >100 chars.low — Only commit subject available, or --max-walk cap was hit.none — Direct push, one-line subject, no recoverable why.Configurable ticket prefixes via TICKET_PREFIXES env var (e.g., TICKET_PREFIXES=CAT,COVEX) to prevent ghost tickets from over-permissive regex.
Used internally by /devkit:pr-review for "why" inferences — this command is the single source of truth for git archaeology.
/devkit:address-pr — Address PR review feedback efficientlyAuthor-side companion to /devkit:pr-review. Reads open reviewer comments, classifies them, drafts code fixes for change-requests and replies for questions, commits in smart batches, posts replies, resolves threads, and re-requests review — all with explicit author approval at every step.
Usage:
/devkit:address-pr 409
/devkit:address-pr 409 --dry-run # show plan only, no changes
/devkit:address-pr 409 --ignore-bots # skip CodeRabbit / dependabot etc.
/devkit:address-pr 409 --reviewer=senior-rev # only this reviewer's comments
/devkit:address-pr 409 --auto-resolve # mark threads resolved after fixes land
Comment classification:
| Type | Action |
|---|---|
| change-request | Draft code fix; ask before applying |
| nit | Batch with other nits in the same file → one commit |
| question | Draft reply; queue for batch posting |
| suggestion | Draft both a fix proposal and a discussion reply |
| praise | Acknowledge in summary, no action |
| out-of-scope | Reply acknowledging deferral; do not auto-resolve |
| stale | Detect line already changed; suggest resolving without code change |
| conflict | Surface to author when two reviewers ask for opposing changes — never silently picks one |
Smart features:
The author stays in control. Nothing is committed or posted without explicit confirmation.
npx claudepluginhub shiyas1331/devkit --plugin devkitComprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Complete developer toolkit for Claude Code
Intelligent draw.io diagramming plugin with AI-powered diagram generation, multi-platform embedding (GitHub, Confluence, Azure DevOps, Notion, Teams, Harness), conditional formatting, live data binding, and MCP server integration for programmatic diagram creation and management.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.