What's Changed

• launch: keep Claude Code channels available by @hoyyeva in https://github.com/ollama/ollama/pull/17210
• cmd: remove dead agent prompt wrappers by @ParthSareen in https://github.com/ollama/ollama/pull/17227
• agent: reorder working directory instruction by @ParthSareen in https://github.com/ollama/ollama/pull/17228
• agent: clean up semantics, UX, DX, and procedural code by @ParthSareen in https://github.com/ollama/ollama/pull/17212
• agent: allow unlimited tool rounds for cloud models by default by @ParthSareen in https://github.com/ollama/ollama/pull/17217
• anthropic: close text block before starting thinking block by @hoyyeva in https://github.com/ollama/ollama/pull/17225
• agent: skills system by @ParthSareen in https://github.com/ollama/ollama/pull/17203
• llama.cpp update by @dhiltgen in https://github.com/ollama/ollama/pull/17186
• cmd: route root command server start through checkServerHeartbeat by @ParthSareen in https://github.com/ollama/ollama/pull/17245
• agent/tui: remove redundant context-window refreshes from event loop by @ParthSareen in https://github.com/ollama/ollama/pull/17241
• cmd: complete slash commands before submitting by @ParthSareen in https://github.com/ollama/ollama/pull/17230
• cmd: remove standalone agent command by @ParthSareen in https://github.com/ollama/ollama/pull/17229
• cmd/tui/chat: render bold emphasis consistently across markdown by @ParthSareen in https://github.com/ollama/ollama/pull/17224
• launch: update Hermes integration by @ParthSareen in https://github.com/ollama/ollama/pull/17202
• MLX update by @dhiltgen in https://github.com/ollama/ollama/pull/17189
• build: bump Linux toolchain to GCC 13 by @dhiltgen in https://github.com/ollama/ollama/pull/17244
• cuda: add CC 10.0 for linux in CUDA v12 by @dhiltgen in https://github.com/ollama/ollama/pull/17025

Full Changelog: https://github.com/ollama/ollama/compare/v0.32.1...v0.32.2-rc0