Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/davidtoby/agent-skillsnpx agentmods add skills/davidtoby/agent-skills/litellm-vertex-proxy-repairWrote 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/davidtoby/agent-skills/litellm-vertex-proxy-repair)<a href="https://agentmods.dev/skills/davidtoby/agent-skills/litellm-vertex-proxy-repair"><img src="https://agentmods.dev/badge/skills/davidtoby/agent-skills/litellm-vertex-proxy-repair/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/davidtoby/agent-skills/litellm-vertex-proxy-repair"><img src="https://agentmods.dev/badge/skills/davidtoby/agent-skills/litellm-vertex-proxy-repair.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00102 | $0.02991 |
| Opus 5 | $0.00051 | $0.01496 |
| Sonnet 5 | $0.00020 | $0.00598 |
| Haiku 4.5 | $0.00010 | $0.00299 |
Grade A, and why
litellm-vertex-proxy-repair scanned grade A with 1 finding 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 8d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- `curl http://127.0.0.1:4000/` fails How it starts
The opening of the file, as written. The whole thing — 334 lines — stays where its author put it; the contents beside it link to each section on GitHub.
LiteLLM Vertex Proxy Repair
A battle-tested repair skill for Toby-style local LiteLLM deployments on macOS.
Use this when a LiteLLM proxy backed by Vertex AI stops serving http://127.0.0.1:4000/v1, when Admin UI fails, or when you need to separate API recovery from PostgreSQL/Prisma recovery.
Quick start
- Confirm whether the proxy is actually listening on port
4000. - Distinguish
API-only outagefromfull-mode DB/UI outage. - Check whether Python is inheriting a macOS system proxy such as
http://127.0.0.1:1082. - If Prisma or Admin UI is involved, force loopback bypass with
NO_PROXY=127.0.0.1,localhost,::1. - Split the service into
liteandfullmodes so API recovery does not depend on PostgreSQL/Prisma/Admin UI. - Verify
/,/v1/models, and optionally/ui/login/separately.
Deep Dive: API-Only vs Full DB Mode & Prisma Migrations (双语解析)
The Problem (问题背景)
English: LiteLLM running in "API-only" mode lacks database configuration (DATABASE_URL). While the /ui/login/ page might still load, submitting a login crashes the backend with Authentication Error, Not connected to DB! because session and user validation require a database. Furthermore, simply connecting a database isn't enough; if Prisma baseline migrations are skipped or interrupted, the backend background jobs will crash with errors like relation "LiteLLM_SpendLogs" does not exist.
中文: LiteLLM 在“纯 API 代理”模式下运行时，没有配置数据库（DATABASE_URL）。虽然你可以打开 /ui/login/ 页面，但在提交登录时，后端由于需要校验用户和生成 session，会直接崩溃并提示 Not connected to DB!。此外，仅仅连上数据库是不够的；如果跳过或中断了 Prisma 的基线同步（Baseline Migration），后端的统计进程就会不断报错崩溃，例如提示 relation "LiteLLM_SpendLogs" does not exist（缺少对应的表或视图）。
The Repair Strategy (修复策略)
English:
- Database Setup: Install and run local PostgreSQL. Inject
DATABASE_URLandDIRECT_URLinto.env. - Prisma Dependencies: Reinstall LiteLLM with proxy extras (
uv tool install --reinstall 'litellm[proxy,extra-proxy]') and executeprisma generateto build the local ORM client. - Lite vs. Full Mode Architecture: Implement a fallback toggle (
LITELLM_MODE).fullenables the DB and UI console.litestrips DB variables on startup, reverting to a pure API proxy to ensure AI generation stays online even if the database fails.
- Full Schema Migration: Run Prisma baseline migrations to synchronize all 100+ tables and views (including
LiteLLM_SpendLogs). This resolves the missing relation crashes and fully restores the UI backend.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 8d ago First seen · 334 lines · 102 tokens per session scan A 010f874d4fbf
litellm-vertex-proxy-repair is a skill published in the GitHub repository davidtoby/agent-skills (10 stars, last pushed 1mo ago), licensed MIT. It adds 102 tokens to every session and 2,991 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
gh-issues
Fetch GitHub issues, spawn sub-agents to implement fixes and open PRs, then monitor and address PR review comments. Usage: /gh-issues [owner/repo] [--label bug] [--limit 5] [--milestone v1.0] [--assignee @me] [--fork user/repo] [--watch] [--interval 5] [--reviews-only] [--cron] [--dry-run] [--model glm-5]…
mistake-analyzer
Analyze incorrect answers from text, documents, or images; identify root causes of mistakes; classify error types; explain correct reasoning; and generate targeted improvement suggestions, wrong-question notebook entries, or similar practice questions. Use when the user provides a wrong answer, asks why they were…
ehr-analysis
End-to-end EHR predictive modeling pipeline with PyHealth, covering dataset loading, task definition, model training, evaluation, calibration, and clinical interpretation.
bindcraft
End-to-end binder design using BindCraft hallucination. Use this skill when: (1) Designing protein binders with built-in AF2 validation, (2) Running production-quality binder campaigns, (3) Using different design protocols (fast, default, slow), (4) Need joint backbone and sequence optimization, (5) Want high…
esm2-sequence-scoring
ESM2 protein language model for sequence scoring, embeddings, and plausibility checks. Use this skill when: (1) Computing pseudo-log-likelihood (PLL) scores, (2) Getting protein embeddings for clustering, (3) Filtering designs by sequence plausibility, (4) Zero-shot variant effect prediction, (5) Analyzing…
scrna-preprocessing-clustering
Standard scRNA-seq preprocessing and clustering with Scanpy. Use for QC, normalization, HVG selection, PCA, neighbor graph construction, UMAP, Leiden clustering, and export of an analysis-ready AnnData object.