v0.10.2

What's Changed

Chat Turbocharging

The chat engine got a caffeine shot: messages are now stacked line‑by‑line, and you can keep typing while the agent thinks. Even the prompt echo now politely says “You: ”, so you never wonder what you just typed. (feat/chat, fix/chat)

CLI Superpowers

The command‑line UI is now a sleek cockpit. Live panels are bounded for reasoning and sub‑agents, buffering adapts on‑the‑fly for tables and code blocks, and streaming deltas are throttled to a buttery ~80 ms cadence. We also added a blank line before metrics, polished the reasoning delimiter, and gave you optional agent description masks for API keys. (feat(cli), perf(cli), fix(cli))

LLM Provider Magic

You can now run Llama CPP locally! A brand‑new provider with configurable base URLs lands, plus client‑side parsing of reasoning tags for local models. The docs got a makeover to explain the new llama.cpp support. (feat(llm), docs)

Workflow Resilience

Scheduled runs that crash are finally recorded and surfaced on startup, so you’ll know exactly which cron job went kaput without digging through logs. (fix(workflow))

CI & Automation Boost

Continuous integration now hands Jazz agents PR context, fixes the workspace path for the PR‑assistant, uses fully‑namespaced model IDs for OpenRouter, and even ships a secret key for OpenRouter API in the release script. (feat(ci), fix(ci), chore(release))

Housekeeping

We pruned a bunch of dead files and unused dependencies (like @effect/workflow, ink‑text‑input, fs‑extra, etc.) to keep the repo lean and fast. (chore)

---

Commits

• ad57363 feat(chat): stack queued messages, one entry per line (#220) by @Landry Monga
• 9ba9840 feat(cli): bounded live panels for reasoning and subagents (#213) by @Landry Monga
• b674518 fix(chat): echo "You: " when draining the queued message (#217) by @Landry Monga
• fb4f804 chore: prune dead files and unused dependencies (#219) by @Landry Monga
• 17fdcec perf(cli): adaptive buffering for tables and code blocks (#218) by @Landry Monga
• d1150b4 feat(chat): queue messages typed while agent is busy (#212) by @Landry Monga
• 91972f7 fix(workflow): record failed scheduled runs and surface them on startup (#215) by @Landry Monga
• 861bcf2 perf(cli): buffer streaming deltas to ~80ms cadence (#216) by @Landry Monga
• 859929c fix(cli): rendering polish — blank line before metrics + reasoning delimiter (#211) by @Landry Monga
• e480f06 fix(cli): polish input flows — optional agent description + mask API key echo (#210) by @Landry Monga
• da1f4cc fix(cli): rebuild streaming pipeline on single-pending-buffer model (#209) by @Landry Mongoose
• 855c0af feat(llm): client-side reasoning-tag parsing for local models (#208) by @Landry Monga
• ac686d0 chore(docs): ignore AI agents docs by @lvndry
• bf4d2b2 docs(plan): implementation plan for llama.cpp provider support by @lvndry
• 7c0050f docs(spec): design for llama.cpp provider support by @lvndry
• 63a7bb3 feat(ci): give jazz agents PR context and finish the workspace fix (#207) by @Landry Monga
• c8fedc6 feat(llm): add llama.cpp provider and configurable base URLs (#199) by @Landry Monga
• 000c8b6 fix(ci): give pr-assistant the workspace path and stop dumping raw failures (#205) by @Landry Monga
• 0d9ba9f fix(ci): use full namespaced model id for openrouter agents (#202) by @Landry Monga
• 229cc10 chore(release): OPENROUTER_API_KEY by @lvndry
• f20e237 Audit follow-ups: persona profiles, security, tokenizer, skills bloat, markdown (#198) by @Landry Monga

Full diff

v0.10.1...v0.10.2