By yamchinsky
Architecture review gate: audits a diff against the repository's own documented structural contracts (docs/architecture/ or ARCHITECTURE.md) and returns findings with rule identifiers, verbatim evidence, and a PASS / FAIL / NOT-APPLICABLE verdict.
A Claude Code plugin marketplace for the DevDigest team AI harness. It packages the reusable parts of the harness — the SDD (Spec-Driven Development) workflow, shared engineering skills, a research agent, and a generalized architecture reviewer — so any team repository can install them without copying files from the DevDigest working tree.
| Plugin | What it provides | Depends on |
|---|---|---|
engineering-paved-path | 12 shared knowledge skills (React, Next.js, Fastify, Drizzle, PostgreSQL, Zod, TypeScript, security, onion architecture, Mermaid, testing) | — |
research-tools | researcher — generic read-only research agent (codebase + web) | — |
architecture-review | architecture-reviewer — audits diffs against the repository's own documented architecture rules | engineering-paved-path |
sdd-engineering | The SDD workflow: spec-creator, implementation-planner, implementer, plan-verifier agents; run-plan, workflow-retro, engineering-insights skills | all three above |
The DevDigest harness files were classified into four groups. Only the first group ships here.
| Group | Components | Disposition |
|---|---|---|
| Reusable | 12 knowledge skills (React/Next/Fastify/Drizzle/PostgreSQL/Zod/TypeScript/security/onion-architecture/Mermaid/testing); researcher; architecture-reviewer (generalized); spec-creator, implementation-planner, implementer, plan-verifier; run-plan (formerly impl), workflow-retro, engineering-insights (generalized); their references and eval scenarios | Extracted into the four plugins after an editorial pass that removed every DevDigest-specific path, module name, and bare-name reference |
| Project-specific | DevDigest CLAUDE.md, pr-self-review skill + hooks, architecture-reviewer-lite, doc-writer, dependency-checker, the vitest eval harness, the devdigest MCP server, product specs | Stay in the DevDigest repository — they encode its modules, contracts, and CI |
| Optional integrations | Convention MCP tools used by spec-creator for grounding | Removed from required tools:; the agent falls back to CLAUDE.md/docs/code search when absent |
| Local leftovers | Caches, personal memory, experiment workspaces, absolute paths | Not extracted |
Why each reusable component updates together with the SDD workflow: the
workflow is the consumer contract — run-plan dispatches
sdd-engineering:implementer and the review gates, the planner assigns
engineering-paved-path:* skills to tasks, and the spec/plan formats are the
interfaces between the agents. Components with an independent consumer
scenario (researcher, architecture-reviewer, the knowledge skills) live in
their own plugins and are consumed as version-constrained dependencies;
workflow-retro has no consumer scenario outside the SDD lifecycle, so it
stays inside sdd-engineering and is invoked manually only.
claude plugin marketplace add yamchinsky/dev-digest-ai-marketplace
claude plugin install sdd-engineering@dev-digest-ai-marketplace --scope project
Installing sdd-engineering automatically installs its three dependencies;
the install output lists them at the end.
A searchable static catalog is published on GitHub Pages:
https://yamchinsky.github.io/dev-digest-ai-marketplace/. It is generated
from the repository files (scripts/build-index.mjs) — never edited by hand.
.claude-plugin/marketplace.json # the catalog manifest
plugins/<name>/ # one directory per plugin (source of truth)
docs/ # guidelines, security, releases, site spec
scripts/build-index.mjs # generates the site search index
site/ # static catalog UI (Vite + React)
Plugin versions live in each plugin's plugin.json. A marketplace commit or
tag pins the state of the catalog; releases are immutable git tags of the form
{plugin-name}--v{version}.
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.
npx claudepluginhub yamchinsky/dev-digest-ai-marketplace --plugin architecture-reviewGeneric read-only research agent: looks things up in the current codebase (with path:line evidence) or on the web (with source URLs), reports in a strict structure, and asks clarifying questions when the request is ambiguous.
Twelve shared engineering knowledge skills: React, React Testing Library, Next.js, frontend architecture, Fastify, onion architecture, Drizzle ORM, PostgreSQL table design, Zod, TypeScript, web security (OWASP), and Mermaid diagrams.
The complete Spec-Driven Development workflow: spec-creator, implementation-planner, implementer, and plan-verifier agents plus the run-plan orchestrator, manual workflow-retro, and engineering-insights capture — wired to shared knowledge skills, a research agent, and an architecture-review gate via plugin dependencies.
Comprehensive C4 architecture documentation workflow with bottom-up code analysis, component synthesis, container mapping, and context diagram generation
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
v9.54.1 — 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.
Documentation generation, code explanation, and technical writing with automated doc generation and tutorial creation
AI-powered wiki generator for code repositories. Generates comprehensive, Mermaid-rich documentation with dark-mode VitePress sites, onboarding guides, deep research, and source citations. Inspired by OpenDeepWiki and deepwiki-open.
Claude + Obsidian knowledge companion. Sets up a persistent, compounding wiki vault (Karpathy's LLM Wiki pattern). v1.7 "Compound Vault" + v1.8 methodology modes close 5 of 5 priority gaps from the May 2026 compass artifact. Ships: substrate alignment with kepano/obsidian-skills, default Obsidian CLI transport, hybrid retrieval (contextual prefix + BM25 + cosine rerank per Anthropic's Sept 2024 research), per-file advisory locking for multi-writer safety, pre-commit verifier agent, AND methodology modes (LYT / PARA / Zettelkasten / Generic) for first-class organizational support no other Claude+Obsidian competitor offers. v1.7.x audit closure: every BLOCKER + HIGH + MEDIUM + LOW finding from the v1.7.0 audit is CLOSED or DEFERRED-with-rationale. Optional DragonScale Memory extension (log folds, deterministic addresses, semantic tiling lint, boundary-first autoresearch).