Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/lonormaly/builders-stack/run-lean-on-neonnpx skills add lonormaly/builders-stack --skill run-lean-on-neongit clone --depth 1 https://github.com/lonormaly/builders-stackWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/lonormaly/builders-stack/run-lean-on-neon)<a href="https://agentmods.dev/skills/lonormaly/builders-stack/run-lean-on-neon"><img src="https://agentmods.dev/badge/skills/lonormaly/builders-stack/run-lean-on-neon.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00080 | $0.01490 |
| Opus 5 | $0.00040 | $0.00745 |
| Sonnet 5 | $0.00016 | $0.00298 |
| Haiku 4.5 | $0.00008 | $0.00149 |
Grade A, and why
run-lean-on-neon scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run lean on Neon
The stack's prod DB is Neon (serverless, managed Postgres). That changes the whole "tune Postgres" story: you do not own a postgresql.conf, so most classic advice (PgTune output — shared_buffers, checkpoint_timeout, max_wal_size, maintenance_work_mem) does not apply — Neon manages it. What you do control is connections, the query/schema shape, a few session GUCs, and how often the app wakes the compute. Those are where "crashes at load" and "blew the free tier" actually come from.
When to use
- Picking a connection string, or the DB falls over under concurrency.
- Choosing how auth/session validation talks to the DB.
- Staying inside Neon's free tier (storage or compute).
Neon free tier — the limits that bind (verify at https://neon.com/pricing, they change)
| Limit | Free | What it means |
|---|---|---|
| Storage | 0.5 GB / project | overflow → writes fail. Don't put blobs in Postgres. |
| Compute | 100 CU-hours / mo (~400 h at 0.25 CU) | overflow → compute suspended till next month. The real bottleneck. |
| Autosuspend | after 5 min idle (can't disable) | your friend — an idle DB costs zero compute. |
| Branches | 10 / project | prune dev branches; change history counts toward storage. |
Compute (CU-hours) — the bottleneck. Don't wake the DB for nothing.
An idle client makes zero requests → compute suspends → CU-hours preserved. Everything that pins compute awake is the enemy:
- Never poll or keepalive the DB. This is the stack's "push, don't poll" law applied to Postgres — a
setIntervalhealth-ping or a per-minute cron holds compute open 24/7 and burns all 400 hours. Use SSE/WebSocket for status, let the DB sleep. - Cache the session cookie — the single biggest saver. Better Auth's
getSessionruns on nearly every request; each one is a DB round-trip that keeps compute awake.session.cookieCache(ON by default inlibs/auth) serves the session from a signed 5-min cookie and only touches the DB on miss. Do not extend the TTL to "save more" — it's a revocation dial, not a cost dial (see the session ladder below). - Batch background work into one wake instead of trickling writes.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 74 lines · 80 tokens per session scan A 9894bfcc1a39
run-lean-on-neon is a skill published in the GitHub repository lonormaly/builders-stack (41 stars, last pushed today), licensed MIT. It adds 80 tokens to every session and 1,490 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
database-and-schema
Database setup, Drizzle ORM schemas, migrations, and the Ops business logic layer.
postgres-drizzle
Proactively apply when creating APIs, backends, or data models. Triggers on PostgreSQL, Postgres, Drizzle, database, schema, tables, columns, indexes, queries, migrations, ORM, relations, joins, transactions, SQL, drizzle-kit, connection pooling, N+1, JSONB, RLS. Use when writing database schemas, queries, migrations…
supabase-postgres-best-practices
Postgres performance optimization and best practices from Supabase. Use this skill when writing, reviewing, or optimizing Postgres queries, schema designs, or database configurations.
Drizzle ORM Testing
Testing patterns for Drizzle ORM covering migration testing, query builder testing, transaction testing, and database integration testing with PostgreSQL, SQLite, and MySQL.
data-table
DiceUI data-table system for building production-ready data tables with server-side filtering, sorting, and pagination. Based on TanStack Table + nuqs URL state + Drizzle ORM backend. Use this skill whenever building, modifying, or debugging data tables — whether adding columns, filters, sort controls, server-side…
setup
Set up memex with provisioned Postgres (RDS + pgvector), AGENTS.md injection, first import.