惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

MyScale Blog
MyScale Blog
Jina AI
Jina AI
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
阮一峰的网络日志
阮一峰的网络日志
P
Proofpoint News Feed
Last Week in AI
Last Week in AI
博客园 - 司徒正美
Martin Fowler
Martin Fowler
T
Tailwind CSS Blog
B
Blog RSS Feed
Vercel News
Vercel News
博客园 - 聂微东
I
InfoQ
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
M
MIT News - Artificial intelligence
Recent Announcements
Recent Announcements
GbyAI
GbyAI
L
LangChain Blog
酷 壳 – CoolShell
酷 壳 – CoolShell
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
Microsoft Security Blog
Microsoft Security Blog
C
Check Point Blog
MongoDB | Blog
MongoDB | Blog
B
Blog

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Flowork Settings: Managing API Keys, Global Defaults, and...
floworkos · 2026-06-16 · via DEV Community

Your owner-level control panel lives here — global stuff that isn't tied to any one agent, kept in the global flowork.db. Every API key, token, and default lives in one place, never hardcoded in code.

Account & Security

The Account section lets you change your password or log out. Keep these credentials secure; they gate everything below.


API Key Management

This is where every external credential goes — your Dev.to key, social cookies, and so on. Nothing is ever written into the source code; agents read these at boot from here.

The easy path: service chips. Under the input there's a row of service chips (Dev.to, X, LinkedIn, Telegram…). Tap one and it auto-fills the exact variable name for you — you don't have to guess whether it's DEVTO_API_KEY or DEVTO_KEY. Then paste the value and hit Save. A chip turns green once its key is set, so you can see at a glance what's already configured.

Manual entry. Type the name yourself (must be UPPER_SNAKE_CASE, e.g. ETHERSCAN_API_KEY) in the first box, the value in the second, then Save. Name suggestions appear as you type.

Edit and delete. Each saved key shows masked — only the last 4 characters visible. Edit re-fills the name and clears the value box for a fresh paste; Delete removes it entirely.

Reserved names are blocked on purpose — anything that could hijack the process or its child commands (PATH, LD_*, DYLD_*, FLOWORK_*, HOME, GIT_*, …). That's a safety rail, not a limitation.

How it reaches agents. When you save a key it's stored in flowork.db and injected into the running process immediately (no restart needed). On the next boot, keys are loaded before agents start, so an agent always sees them in its environment. A dev platform agent, for example, reads its own key via os.Getenv("DEVTO_API_KEY").


Router & Model Configuration

The global default that agents fall back to when they don't pin their own:

  • Default model — e.g. claude-haiku-4-5. Leave it empty to use the built-in default.
  • Router URL — e.g. http://127.0.0.1:2402. Leave it empty for the built-in local router. Must be a localhost address — an external URL is rejected and falls back to the default (a safety rail against routing traffic somewhere it shouldn't go).

Per-agent settings always win. If an agent sets its own model, that choice is kept. These two values only fill the blank for agents that set nothing. Saving takes effect live and is also applied before agents boot.

Why use the local router? It injects the anti-hallucination antibody — your agent's most-recurrent, most-relevant past mistakes ranked by relevance — before the model answers. A raw third-party API gets none of this. The router is model-agnostic; swap the model in one place and route to a subscription service or a local model seamlessly.


Telegram Notifications

Paste your bot token + chat ID, hit Save, then Test (it sends you a test message). This is the token the whole system uses to ping you — yours, never hardcoded. It's how agents (and the system itself) reach you when something matters.


OAuth Integration: YouTube

Connect a YouTube account by OAuth for automations that need it. The credential is stored securely here and available to agents that declare it in their tools list.


Guardian — Integrity Watch

The Guardian shows you which files it's watching for integrity violations. It's your immune system: monitoring the agent's doctrines and memories for poisoned or corrupted data, and quarantining threats before they spread.


One Place, One Principle

Everything global belongs here — never paste a key into a file, a per-agent box, or environment directly. If a tutorial asks you to, it's out of date. Your keys, your defaults, your Telegram token, your OAuth creds — one place, kept apart from each agent's private settings. Agents are isolated citizens; they share nothing unless you wire them. Disable one and nothing else notices.

🔗 Open source on GitHub

💬 Join the Flowork community on Telegram: https://t.me/+55oqrk75lc43YWE1