AINewsnow

Cut my always-loaded Claude Code context by 57.5% (7,337 -> 3,121 tokens) and built a model-compatibility lint + skill factory (tq-forge, MIT)

After building dozens of custom skills and subagents in Claude Code, two problems kept biting me: Always-loaded skill definitions were burning 7,300+ tokens before I even typed a prompt. A skill that worked great on Opus would quietly degrade on Sonnet or Haiku—and pinned model IDs in templates kep…

Read the full story at DEV Community — AI ↗

Timeline · 1 report

  1. 2026-11-13 16:52 · DEV Community — AI
    Cut my always-loaded Claude Code context by 57.5% (7,337 -> 3,121 tokens) and built a model-compatibility lint + skill factory (tq-forge, MIT)

More stories

  1. Jensen Huang resigns as Nvidia CEO amid spy scandal and chip smuggling probes — CSET Georgetown
  2. Microsoft 365 outage disrupts Outlook, Teams and Copilot for over a day — Euronews Next
  3. Anthropic expands Claude into workspace suite with Sheets, Whiteboards, Data and Meetings — Anthropic (X)
  4. Anthropic launches Claude Fable 5.2 and Claude Mythos 5.2 — DEV Community — AI
  5. Anthropic launches self-hosted Claude Managed Agents in private beta — Anthropic (X)
  6. Anthropic launches Claude Fable 5.2 and Mythos 5.2 as model race intensifies — AWS Machine Learning Blog
  7. Anthropic prepares IPO at valuation below once-touted $2 trillion — Reuters AI
  8. Mariko Mori uses Claude to create animation for retrospective 'All That Shines' — Anthropic (X)

Get the daily brief of stories like this at 6:30 every morning →