implementing #codex already spawning via structured DELEGATE phew codex --profile agentic-engineers-orchestrator --sandbox workspace-write --ask-for-approval on-request

it’s the Delegate Handback Protocol Revival surprisingly nebulous, but I’m STILL not running out of tokens!! Multiple layers, all compounding but just you wait I’ve glimpsed beyond

GitHub

Comparing v0.40.0...v0.40.1 · niallyoung/agentic-engineers

portable sub-agent orchestration "AGENTS with SKILLS" (Orchestrator delegates 1..N specialist AGENTS (model+skill/s) optimising for token...

holy crap disappearing events forget more users, get the basics right - implement the actual protocols!!

#NOSTR devs and engineers get in first please 🙏 I CUT MY TOKEN USAGE IN HALF $$$ all while quality went up, quicker delivery sequential or parallel, pick your harness and provider WIP and a bit of public release and tag cleanup just now 🤦‍♂️ but hey that’s a new skill ♻️ 😁

GitHub

GitHub - niallyoung/agentic-engineers: portable sub-agent orchestration "AGENTS with SKILLS" (Orchestrator delegates 1..N specialist AGENTS (model+skill/s) optimising for token usage, quality and speed ♻️self-improving♻️ `make install` ~/ configs for CLI harnesses: Claude, Copilot, Codex, OpenCode, Pi

portable sub-agent orchestration "AGENTS with SKILLS" (Orchestrator delegates 1..N specialist AGENTS (model+skill/s) optimising for token...

announcing Agentic Engineers — an open-source, production-ready multi-agent orchestration framework that solves co-ordinating agents in parallel, across multiple LLM providers and Agent Harnesses. OpenCode, Copilot CLI, Claude Code, π.dev 👉 GitHub:

GitHub

GitHub - niallyoung/agentic-engineers: portable sub-agent orchestration "AGENTS with SKILLS" (Orchestrator delegates 1..N specialist AGENTS (model+skill/s) optimising for token usage, quality and speed ♻️self-improving♻️ `make install` ~/ configs for CLI harnesses: Claude, Copilot, Codex, OpenCode, Pi

portable sub-agent orchestration "AGENTS with SKILLS" (Orchestrator delegates 1..N specialist AGENTS (model+skill/s) optimising for token...

Key Benefits & Discoveries DELEGATE/HANDBACK Protocol Enforces Quality — 90+/100 quality, 40-60% faster, 80% fewer iterations Token Efficiency: 40-60% Reduction via Smart Model Selection — 65% cost savings vs. all-Opus Parallel Sub-Agent Execution at Scale — tens to hundreds of concurrent agents, 5-tier hierarchies Use your preferred Harness Support My Work This framework is the result of months of experimentation and hard work. I'm bootstrapping this independently. If Agentic Engineers saves you money or time, please consider supporting me to keep going! #AI #MultiAgent #OpenSource #SoftwareEngineering #AgenticAI #Orchestration #LLM #Model #Harness #claude #copilot #opencode

Well done Kraken 👍 `sign a non-transactional message to confirm you control the wallet's private key` ``` What's changing: Transactions involving private wallets will require a one-time wallet ownership verification. Once verified, that wallet address won't need verification again. Two simple ways to verify: - Self-Certification: one-click confirmation + Funding 2FA check. - Digital Signature: sign a non-transactional message to confirm you control the wallet's private key. Learn more about Digital Signature here. ```

copilot cli FTW

I authored FAI-BootCD … 2002? 2003? linux.conf.au 2003 live demo install of a freshly installed and configured Debian server in about 90 seconds. Re-image? just `rm /boot/vmlinuz; shutdown -r now` and grub default reinstall FAI on ***READ-ONLY BOOT MEDIA*** and boom there was our Corporate Dev, Test and Prod server build infra with verifiable sources, checksums, known builds and custom packages from our own infra, just manage updates 🫡 I think it probably lasted near to the end, another decade at least @ ii 🤷‍♂️ need that shit again but optical disappearing 😱

if you think about it we’re all PERL devs now Pathologically Eclectic Rubbish Lister literally what LLMs do 🤣 auto-complete at scale lisp is also nice, but perl got it done mid 90s 🫡

?? 🍿

58k incoming?