RepoReview — MelTuc

// DASHBOARD

Review Command Center

The dashboard gives you an instant snapshot of your content pipeline. Six stat tiles show total posts published, drafts awaiting review, dismissed entries, unreviewed repos still waiting in GHT, reviews generated today, and next scheduled run. A status strip breaks down draft/published/dismissed distribution, and the recent publications list shows the last five reviews sent live to meltuc.com/blog. Last generation metadata surfaces the cron run time, tokens used, and any skip reasons from the quality gate.

meltuc.tech/rrv/app

128

Published

Drafts

Dismissed

169

Unreviewed

Today

Auto-Pub

Draft Status Distribution

Published 128

Draft 7

Dismissed 34

Publication History (12 weeks)

▶ Publish Next

Recent Publications

Title Repo Lang Stars Published

Why browser-use is the missing agent primitivebrowser-usePython42kToday

Zed editor finally ships collaborative AIzedRust48kYesterday

Bun 1.2 benchmarks: is Node officially obsolete?bunZig72k2d ago

Inside the Astro 5 content layer rewriteastroTS44k3d ago

Deno 2 and the great runtime reconciliationdenoRust94k4d ago

Last run: 06:30 UTC · Next: in 4h 12m · 3/3 drafts passed quality gate

// DRAFT MANAGEMENT

Editorial Control Grid

Every generated review lands in a sortable data grid showing title, repo slug, primary language, star count, trend status, word count, and publication state. A filter bar at the top lets you search by keyword, filter by status (draft, published, dismissed), or language. Clicking any row slides out a reader panel on the right with the full rendered markdown, source repo metadata, and one-click actions: edit, publish, dismiss, or delete & reset so the repo becomes available again for a fresh take.

meltuc.tech/rrv/app/posts

Search drafts...

All Status

All Languages

Sort: Newest

Title Lang Stars Words Status

Why browser-use is the missing agent primitivePython42k1240Published

Uv, the Python package manager that finally clicksRust28k980Draft

Notes on the shadcn/ui explosionTS68k1150Draft

Zed editor finally ships collaborative AIRust48k1320Published

Bun 1.2 benchmarks: is Node obsolete?Zig72k1410Published

Yet another yaml config toolkitGo2k410Dismissed

Astro 5 content layer deep diveTS44k1080Published

DRAFT PREVIEW

Why browser-use is the missing agent primitive

Repo Metadata

browser-use/browser-use
Python · MIT · 42k stars
+1.2k stars this week

Quality Checks

Word count1240 ok

Structurepass

Repo mentionpass

Tokens8,412

Actions

Edit Publish Dismiss Delete

// GENERATION LOG

Full Audit Trail

Every generation run is logged with the repo slug, trigger (cron or manual), duration, token usage, quality check result, and skip reason if the draft was rejected. The log makes it easy to diagnose why a repo failed the gate, see which cron runs produced which published posts, and track API cost over time. Manual regeneration of any repo is one click away from here.

meltuc.tech/rrv/app/log

GENERATION RUNS (last 7 days)

All results

When Trigger Repo Tokens Duration Result

06:30 todaycronbrowser-use/browser-use841254sPublished

06:30 todaycronastral-sh/uv792061sDraft

06:30 todaycronshadcn-ui/ui810258sDraft

14:22 ydaymanualzed-industries/zed910472sPublished

06:30 ydaycronrandom/yaml-config281038sSkip: word count

06:30 ydaycronoven-sh/bun864062sPublished

06:30 2dcronsome/empty-readme4129sSkip: no README

7d tokens: 214,308 · avg duration: 52s · gate pass rate: 78%

// SETTINGS

Configuration

Settings shape how RRV writes. Toggle auto-publish on or off, set the maximum number of drafts generated per cron run, pick the review tone (analytical, enthusiastic, skeptical), define the word count target, and control the quality gate cutoffs. Source controls let you weight GHT trending repos versus STT trending skills as the pool from which tomorrow's review is drawn.

meltuc.tech/rrv/app/settings

Publishing

Auto-Publish

Enabled

Max Posts Per Run

Cron Schedule

Daily 06:30 UTC

Review Voice

Tone

Analytical / opinionated

Target Words

1100 – 1400

Source Weights

GHT Repos

70%

STT Skills

30%

Quality Gate

Min Words

300

Max Words

5000

Require Repo Mention

Yes

Blog Endpoint

meltuc.com/blog

// HOW IT WORKS

From Trending Repo to Published Review in 5 Steps

Every published post walks the same automated pipeline — from candidate selection to meltuc.com/blog.

Pick Candidate

Pull unreviewed repos from GHT's trending tracker and hot skills from STT. Weight by star velocity, recency, and source mix.

Fetch Metadata

Hit the GitHub API for README, recent commits, top contributors, dependency files, releases, and language breakdown.

Claude Drafts

Structured prompt with full repo context feeds Claude Sonnet. Output is an honest, opinionated developer review in markdown.

Quality Gate

Validates word count, markdown structure, and repo mention. Failures are logged with a skip reason; the repo stays eligible.

Auto-Publish

Best surviving draft is promoted to meltuc.com/blog. Remaining drafts queue for editor review or the drip publisher.

// WHAT YOU GET

Everything You Need to Run a Content Pipeline

RRV is built on focused subsystems that each do one job and all plug into the same daily run loop.

🔗

GHT Integration

Reads directly from GitHub Trends Tracker. No duplicate discovery — RRV uses the repo pool GHT has already scored and deduped by star velocity.

📡

STT Integration

Pulls hot skills from Skills Trends Tracker so reviews can theme around the technologies developers are actively searching for this week.

🤖

Claude Reviews

Claude Sonnet drafts every review from a structured prompt with full repo context. Output is opinionated long-form markdown, not a README summary.

📦

GitHub Metadata

README, recent commits, top contributors, dependency files, release notes, and language breakdown — all fetched fresh for every generation run.

⚡

Auto-Publish

Daily 06:30 UTC cron generates drafts and promotes the best one to meltuc.com/blog. Consistent cadence, zero manual effort.

📝

Draft Management

Edit, publish, dismiss, or delete & reset any draft. Deleting a bad review frees the repo for a fresh take — the system always recovers gracefully.

📊

Publication History Chart

A 12-week bar chart shows weekly publication output at a glance. One-click "Publish Next" drains the draft queue without touching individual posts.

🔍

SEO Keyword Extraction

At publish time, 8 SEO keywords are automatically extracted from the post body and stored as tags — ready for meta tags and structured data.

📰

Newsletter Coverage Badge

Posts featured in a NLG newsletter show a badge in the post grid. At a glance you know which content has been promoted to avoid duplicate coverage.

⏰

Drip Publisher

A daily 9am cron auto-publishes one pending draft, adding a second publication slot that keeps the blog cadence steady as the queue builds.

📅

Publication Calendar

A month-at-a-glance API merges scheduled and already-published posts into one calendar view. Plan cadence ahead of time and spot gaps or clusters before they happen.

✨

AI Review Summary

One click generates a sharp 2-sentence summary of any review via Claude Haiku — focused on practical value and who benefits. Saved to the post for reuse in newsletters and blog meta.

📊

Publication History Chart

A 12-week bar chart on the dashboard shows weekly publication output at a glance. Quickly see which weeks were high-output, where the pipeline stalled, and long-term publishing cadence trends.

12-Week View Bar Chart Publish Next

TRENDING REPOS.
DAILY REVIEWS.

Review Command Center

Editorial Control Grid

Full Audit Trail

Configuration

From Trending Repo to Published Review in 5 Steps

Everything You Need to Run a Content Pipeline

Stop Writing. Start Publishing.

TRENDING REPOS.DAILY REVIEWS.

Review Command Center

Editorial Control Grid

Full Audit Trail

Configuration

From Trending Repo to Published Review in 5 Steps

Everything You Need to Run a Content Pipeline

Stop Writing. Start Publishing.

TRENDING REPOS.
DAILY REVIEWS.