A developer's Claude Code subagent harness routes planning and final review to Fable while delegating execution to Opus, Sonnet and Haiku, after a June workflow script with a single model-less agent()…
A developer's measurement of Claude Code subagent startup costs found the same trivial task costs between 15,746 and 38,401 tokens depending on a single dispatch field, while published figures range f…
A developer published `baton-agents` to npm on September 14, four days after Sophia Myang's September 10 request for a cross-platform agent orchestration harness drew 1,024 bookmarks and 340 replies. …
A developer who hardened a personal AI agent that reads email, files, and the desktop reports that the primary security fix is structural, not model-based: a rule forbidding any single task from combi…
A developer reports that after seven months of running an AI agent with persistent memory, the agent's self-maintained markdown files rotted in five specific ways, including stale progress files, refe…
Developer ucsandman built declick, a tool that compiles MCP servers and other API sources into CLIs, reducing the context-window cost of tool schemas from 236,818 bytes to 58,309 bytes (4.1x smaller) …
An AI agent built on Google's AlphaEvolve pattern beat 10 listed Packomania circle-packing records overnight for a total model spend of $27.72, with improvements ranging from 2.5% to 5.4% above prior …
A developer's Claude Code setup after seven months includes 19 hooks, 36 skills, and rules that became code, with sessions on 151 distinct days across about 140 directories. The setup is mirrored publ…
Practical Systems, a company run by three OpenClaw AI agents on separate machines, can now approve some of its own work without human intervention, according to a company update. The agents—MoltFire (…
A new open-source Python package, Context Handoff Bundle, cuts AI coding agent re-orientation token usage by about 97%, resuming sessions with roughly 640 tokens instead of re-reading source files tha…