Agent Forge

@agentforge.bsky.social

3 Followers 12 Following

Tools & kits for people who run AI agents 🛠️ Makers of The Multi-Agent Operating Kit. Free CLAUDE.md Builder & Linter → agentforge-kit.vercel.app

Launch card for the book 'Your Second Brain in Obsidian' by Waqas Nasir.
If naming a function is hard, the function is probably doing too much. A name that needs "and" or "manager" or "helper" is a confession: you haven't found the single responsibility yet. The struggle to name is the design feedback. Listen to it.
The agents that actually work are boringly narrow. One clear job, a tight set of tools, an obvious done condition. The "do everything" agent looks impressive in a demo and falls apart in production. Scope is the feature. Vague autonomy is the bug.
Three command-line tools that earned a permanent spot in my setup: ripgrep (rg) — search that's stupid fast fzf — fuzzy-find anything, pipe it anywhere jq — slice JSON without writing a script None are flashy. All of them save minutes every single day.
Two hours of uninterrupted focus outproduces a full day of fragments. Context-switching has a tax most people never count: every interruption costs the reload, not just the minute. Guard one real block a day. Notifications off, door shut. Protect the depth.
Shipping something slightly embarrassing beats polishing something nobody's seen. The version in your head is always cleaner than the one users need. You learn what's actually wrong only after it's live. Real feedback starts the day you ship, not the day you finish.
A second brain dies from one thing: capture friction. If saving a note takes more than five seconds, you won't do it when it matters — and the idea is gone. Optimize for the speed of capture first. Organize later, or never. Messy and saved beats tidy and lost.
The habit that quietly made me faster: commit small, commit often. A 12-file commit hides bugs and resists review. A 2-file commit explains itself. When something breaks, small commits are a flashlight. Big ones are a haystack.
Your README's first line is the whole pitch. Most open with "This project is a..." — pure throat-clearing. Write what it does for the reader, in one sentence, before any install steps. If a stranger can't tell whether to keep reading, they won't.
Agent Forge card: never edit code you haven't read. Patching what you assume is there deletes someone's work. Read first, then edit.
Agent Forge card: a fix you didn't re-check is a hope. Re-run the failing case with fresh eyes before claiming fixed.
Agent Forge card: stop planning agents with dates. 'Finish by Friday' invites nonsense; 'next, after X confirms' keeps the chain honest. Sequence beats deadlines.
Agent Forge card: don't ask one agent, ask three from different angles then synthesize. Independent passes catch what single reasoning misses.
Agent Forge card: code that compiles isn't code that works. Verify a fix by what the browser actually renders, not what the source claims.
Agent Forge card: the conductor doesn't pick up the instrument. Hoarding a task means specialists never grow; delegating compounds experience and strengthens the whole system.
Agent Forge quote card: building was never the bottleneck — distribution is, and AI didn't change that.
Agent Forge card: write CLAUDE.md like a prompt, not docs. Imperatives only — Always X, Never Y. If a line wouldn't change a decision, it's noise.
Agent Forge card: give each agent only the tools its job needs. Scope the tools, scope the blast radius.
Agent Forge diagram: summaries up, details down. A conductor box feeds three lead boxes with summary-only; each lead absorbs the detail.
Agent Forge card: a fresh agent session remembers nothing. Write down what you learned and read it before you start — memory turns runs into an operation.
Agent Forge card: why we serialize heavy agents. Too many in parallel took services down; running heavy work one at a time stays stable.
Agent Forge card: the one move that breaks everything in multi-agent setups — tearing down a team while agents still work. Steps: stop the workers, confirm done, then dismantle.
5 anti-patterns that get your CLAUDE.md ignored: 1) The kitchen sink (too long, the model skims). 2) Vague aspirations ('write clean code' = nothing). 3) Polite hedging. 4) Stale rules. 5) Calendar dates.