SkillAgentSearch skills...

vibedollar-skill

vibedollar — find your first customers on Reddit. Give it your product description; get back potential-customer leads (posts + comments with why-it-matches annotations + a demand-validation report), backed by a 470+ indie case library. Remote-hosted, zero setup; BYOK, zero LLM deps.

Install / Use

claude mcp add Big-Pi100 -- npx -y github:Big-Pi100/vibedollar-skill

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

79/100

Supported Platforms

Claude Code
Claude Desktop
Cursor

Tags

Our assessment of vibedollar-skill

vibedollar-skill scores 79/100 on our quality scale, 783rd of 954 AI & Machine Learning skills we index.

Its MCP Server is 11 KB long, well organised into 9 sections with 2 code examples: a thorough specification that gives an agent plenty to work with.

It has 3 GitHub stars, so there is little community track record yet; judge it on its content.

Substance
29/30
Structure
18/20
Description
15/15
Adoption
3/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated 23 days ago, so vibedollar-skill is actively maintained.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 92/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-18. Automated pattern scan on 2026-10-18. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

vibedollar-skill compared with similar skills

All 4 of these similar skills score higher than vibedollar-skill; compare them before choosing.

SkillScoreStarsUpdatedFormat
vibedollar-skill (this skill)by Big-Pi10079323d agoMCP Server
claude-memby thedotmack100101.4ktodayCLAUDE.md
Agent-Reachby Panniantong10098.7k2d agoCLAUDE.md
Understand-Anythingby Egonex-AI10086.8k5d agoCLAUDE.md
headroomby headroomlabs-ai10075.6ktodayCLAUDE.md

Frequently asked questions

How do I install vibedollar-skill?
Run claude mcp add Big-Pi100 -- npx -y github:Big-Pi100/vibedollar-skill. The install tabs above show the steps for each supported agent.
Which AI agents does vibedollar-skill work with?
It is written for Claude Code, Claude Desktop and Cursor, as a MCP Server file. Other agents that read the same format can often use it too.
Is vibedollar-skill safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 92/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is vibedollar-skill still maintained?
The repository was last updated 23 days ago, so vibedollar-skill is actively maintained.

vibedollar — find your first customers on Reddit

The customers you're looking for are already complaining and asking for solutions on Reddit. vibedollar brings them to you.

Give vibedollar your product description and it continuously monitors Reddit for potential-customer leads (posts + comments, with author, full text, and why-it-matches annotations). A lead is billed when you claim it; scoring it with your own LLM is free.

  1. MCP service (mcp.vibedollar.net) — Register, subscribe a product, candidates accumulate in the background, your agent scores and claims them. Zero setup (just a product description).
  2. SKILL — makes vibedollar part of your agent's muscle memory.

What it does for you

| Use case | How | |----------|-----| | Cold-start acquisition | vibe_subscribe → system tracks continuously, leads accumulate, open the post and reach the prospect | | Validate a product idea | Subscribe, then look at accumulated leads: N strong-demand signals → build it | | Ongoing acquisition | vibe_subscribe → vibe_leads (billed on claim) → score → reach out | | Proof for investors/team | Accumulated leads (that you claimed) are ready-made demand validation |

Business model: pay for the leads you claim. A lead is billed the first time you claim it; scoring, re-reading and exporting are free. The server meters claims, so nothing is self-reported and nothing is billed twice.


MCP tools

| Tool | What it does | Cost | |------|-------------|------| | vibe_register / vibe_verify | Two-step signup → API key (email code → key) | Free | | vibe_balance | Check balance / tier / monthly claim allowance and daily cap | Free | | vibe_subscribe | Subscribe a product (describe it; background tracking starts) | Free | | vibe_leads | Claim leads (author / body / system reference score) | Billed per lead claimed | | vibe_submit_score | Score claimed leads with your own judgment: relevant → delivered list; irrelevant → feedback for tuning | Free | | vibe_score_discuss | View/respond to disagreements with system reference score (calibration) | Free | | vibe_list_subs / vibe_unsubscribe | Subscription management | Free | | vibe_delivered / vibe_get_delivered | Delivered leads (follow-up) | Free | | vibe_mark_leads | Mark lead outcome (valid / invalid / contacted) | Free | | vibe_recover_key / vibe_recover_verify | Lost your API key? Recover it via email verification | Free |

Billing: a lead is billed when you first claim it (Free 1,000/mo · Starter 5,000/mo then $5 per 1,000 · Pro 30,000/mo then $3.50 per 1,000). Scoring, re-reading and exporting are free. Daily delivery caps (Free/Starter 1,000, Pro 3,000) carry over to the next day.


Quick start

  1. Connect (remote-hosted, no local setup): add endpoint https://mcp.vibedollar.net/mcp to your MCP client (FastMCP HTTP transport).
  2. Register (two-step):
    vibe_register(email="you@example.com")     # sends a 6-digit code
    vibe_verify(email="you@example.com", code="123456")   # returns api_key
    
    Your API key is also emailed to you. Unlock lead allowance: subscribe Starter/Pro, or top up the wallet (WeChat ¥1–¥1000 / Creem $1–$200). Free tier already includes 1,000 claimed leads/month.
  3. Configure auth: put the key in your MCP client request header — Authorization: Bearer <key> (or Api-Key: <key> if your client disallows custom Authorization headers).
  4. Check balance/allowance: vibe_balance() (key read from header automatically).
  5. Web self-service also available: https://vibedollar.net/account.html (register / verify / pay).

Recommended workflow

vibe_subscribe(product="team wiki tool for small teams")
    → {"ok": true, "subscription_id": 12, "message": "tracking started..."}
vibe_list_subs()                          # check status + pending count
vibe_leads(subscription_id=12, limit=10)  # claims 10 leads (billed per lead)
    → {"id": 1, "title": "...", "url": "...", "score": 90, "reason": "...", ...,
       "billing": {"claimed_batch": 10, "free_items": 10, "currency": "usd", "charge": 0.0, "charge_usd": 0.0,
                   "claimed_this_month": 10, "claimed_quota": 5000,
                   "daily_used": 10, "daily_cap": 1000}}
vibe_submit_score(scores=[
    {"id": 1, "verdict": "relevant",   "score": 90, "reason": "directly asking for a solution"},
    {"id": 2, "verdict": "irrelevant", "score": 10, "reason": "unrelated"},
])
    → {"ok": true, "passed": 1, "rejected": 1, "quota_used": 1, "quota_limit": 3000}  # scoring is free
      # BATCH IT: put the whole batch here (up to 100 items) -> N verdicts = ONE pipeline call (1 token).

Rules:

  • A lead is billed when you claim it (vibe_leads returns a billing block: batch, free/billable split, charge, month usage, daily cap)
  • Scoring is free: verdict="relevant" moves the lead to your delivered list (vibe_delivered); irrelevant feeds back to sharpen matching
  • Please also return irrelevant — that's how the system learns your judgment standard
  • Each lead can be scored only once (re-scoring is rejected)
  • The score field is a system reference score — your judgment wins
  • Daily cap: Free/Starter 1,000, Pro 3,000 claimed leads/day; beyond the cap delivery resumes automatically next day (nothing lost)
  • Unprocessed claimed leads come back with ids (never lock your subscription); leads older than 7 days auto-expire
  • Someone else may be judging in parallel: the web app (https://vibedollar.net/app.html → To score) claims and judges on the same tools and the same gate — manual verdicts arrive as score: 60 (relevant) / 1 (irrelevant), whoever claims first judges, and a lead already judged there is rejected as already-scored, so skip that id instead of retrying. The pending queue (50 unscored) is shared between your agent and the web app: either end scoring it unlocks the next batch.

Pricing

| Tier | Price | Claimed leads / month (billed on first claim) | Daily cap | Rate limit (write / pipeline / read) | |------|-------|---------------------------------------------------|-----------|------------| | Free | $0 (sign-up) | 1,000/mo — use them as fast as you like | 1,000/day | 15 / 30 / 120 req/min | | Starter | $19/mo | 5,000/mo, then $5 per 1,000 extra | 1,000/day | 30 / 60 / 240 req/min | | Pro | $79/mo | 30,000/mo, then $3.50 per 1,000 extra | 3,000/day | 60 / 120 / 480 req/min | | Wallet top-up | Any amount (Creem $1–$200 / WeChat ¥1–¥1000) | Covers leads beyond your monthly allowance | — | — |

  • Billing unit = a lead you claim: 1 post or 1 comment = 1 lead (both author and commenter are prospects), billed the first time you claim it. Scoring, re-reading and exporting are free — the server meters claims, so there is nothing to self-report.
  • Daily delivery cap: Free/Starter 1,000 leads/day, Pro 3,000/day. Anything beyond the cap carries over to the next day — nothing is lost and nothing is billed twice.
  • Beyond the monthly allowance: extra leads deduct $5/1,000 (Starter) or $3.50/1,000 (Pro) from wallet credit; insufficient wallet pauses delivery and the leads stay reserved. Free has no overage — upgrade when the 1,000 is used up.
  • Sign-up gives 1,000 free leads/month (1,000/day) — enough to run the full subscribe → claim → score → deliver loop.
  • No usage-inducement: how many subreddits or keywords you track is your call; a bigger plan is never needed just to unlock more.
  • Upgrade: Creem (global) or WeChat Pay (CN) at mcp.vibedollar.net — auto-activation after payment. Annual 20% off (coming soon). No trial.

Payment & activation (agent workflow)

The only user action is scanning a QR code / clicking a link. You (the agent) do the rest:

  1. Confirm/register api_key: vibe_balance() → if "Missing API key", register via vibe_register → user gives you the emailed 6-digit code → vibe_verify. Don't ask the user to self-register.
  2. Pick channel by user location: | Location | Channel | Price | |----------|---------|-------| | Global (default) | Creem (USD card) | Starter $19/mo / Pro $79/mo / Top-up $1–$200 | | Mainland China | WeChat Pay (CNY) | Starter ¥99/mo / Pro ¥399/mo / Top-up ¥1–¥1000 |
  3. Generate payment entry:
    • Creem: POST https://mcp.vibedollar.net/creem/checkout {"api_key": "<key>", "plan": "starter|pro|topup", "email": "..."} → {"url": "<payment link>"} → send the url. Never put api_key in URLs.
    • WeChat: POST https://mcp.vibedollar.net/pay/native {"api_key": "<key>", "email": "...", "tier": "starter|pro|topup"} → code_url → render as QR → user scans. For top-up add "amount_cents": <CNY×100>.
  4. Confirm activation: poll GET https://mcp.vibedollar.net/pay/orders/<out_trade_no> or re-check vibe_balance() — tier/credit auto-updates via webhook/callback.

Troubleshooting:

  • User already has a key → use it directly, don't re-register
  • User prefers web → point them to https://vibedollar.net/account.html (equivalent flow)
  • Tier not updated after payment → wait 1–2s and re-check vibe_balance(); if still stale, report out_trade_no and contact support@vibedollar.net
  • Cancel: Starter/Pro downgrade at expiry (no hard cut), no agent action needed

MCP client config

{
  "mcpServers": {
    "vibedollar": {
      "type": "http",
      "url": "https://mcp.vibedollar.net/mcp",
      "headers": {
        "Authorization": "Bearer <your-key>"
      }
    }
  }
}

Some clients (Claude Desktop / Cursor) also support headers via env/config — set Authorization: Bearer <key> on the server's headers per your client's docs. Alternative header: Api-Key: <key>.

Connect → vibe_register → put the key in the header → call data tools. We handle lead discovery and optimization; you handle judgment.


Status (2026-09-05)

  • Claim-based billing (2026-09-10): a lead is billed when first claimed; scoring, re-reading and exporting are free (server-metered, nothing self-reported)
  • Data-service architecture: server = pure data/state layer (collection + matching + SQL state); judging & keyword tuning are yours (your LLM)
  • 2026-09-05 tools expansion: 9 management tools added — keywords/keyword_add/keyword_remove (word management), sd_update (judgement scope), sub_health/opt_log (delivery-health-driven tuning), rejected/recover_lead (recycle), subs (source stats)
  • 2026-09-06 keyword_add_batch: vibe_keyword_add_batch adds N keywords in one call (same per-word semantics as keyword_add) — use it for init/expand lists instead of looping, to stay inside the per-account rate window
  • Delivery-health loop: read vibe_sub_health → expand/retire keywords on commitment gap → record with vibe_opt_log (replaces old "system auto-tunes")
  • Recycle flow: irrelevant lands in vibe_rejected; recover & re-score with vibe_recover_lead (no double billing)
  • Abuse protection: batch unlocking + claim validation + consistency guardrails

Related Skills

View on GitHub
GitHub Stars3
CategoryAI
Updated23d ago
Forks0

Languages

Python

Trust signals

92/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 low