fix: bash command validation - catch JSON/Python/bare URLs before execution, prevent exit 127 confusion

CRITICAL FIX: SendUserMessage no longer breaks agent loop + match original behavior

CRITICAL FIX: strip ALL custom agent heuristics, return to original simplicity

fix: solution-oriented error handling + less aggressive loop detection

fix: smarter error classification + interleaved text/tool rendering

fix: multi-step agent execution — don't break on text-only when plan has pending steps

fix: ROOT CAUSE of agent infinite loops

fix: 3-layer loop detection — consecutive tool errors, stream error detection, text loop safety

fix: loop detection + MAX_ITERATIONS=200 + consecutive text-only break + revert retry to 2s + openlawX personality

fix: remove prefill self-talk bug + reorder tool cards

feat: openlawX personality, anti-censorship prefill, progress indicators, jump to live

feat: JSON repair, exponential backoff, sidebar date groups, parity audit fixes

feat: complete P0 fixes + Manus UI overhaul phase 2

fix: restore Qwen3 hybrid thinking mode — don't disable it

feat: Qwen3 uncensored prompt + Manus-style UI + P0 bug fixes

HARDCODE DeepInfra naglukho: fix model name, Dockerfile, remove API settings from UI

Fix all 10 critical audit bugs:

feat: DeepInfra Qwen3-Coder-480B-Turbo + provider UI + thinking mode + 30+ model aliases

fix: critical tool chain + planning + autonomy bugs

UNRESTRICTED: Remove all censorship, add jailbreak, maximize autonomy

CRITICAL: Rewrite system prompt for full agent autonomy

429/500/502/503/network: infinite retry every 2s, no backoff, never give up

Fix 429 rate limit: INFINITE retries with exponential backoff (capped 60s)

Fix 7 critical autonomy bugs:

fix: critical compaction, buddy SVG sprites, session events, max_tokens

fix: protect API config from corrupted settings — hardcoded fallbacks + sanitization

feat: auto-compact on context overflow + proactive compaction + token tracking + buddy server events

Full parity with original Rust: session validation, sandbox detection, remote proxy, hooks payload, LSP shutdown

fix: 3 critical bugs to match original Rust implementation

fix: remove 30-step limit — match original usize::MAX (conversation.rs)

fix: revert max_tokens to 16384 (output tokens, not context window)

feat: remove API settings tab, set default max tokens to 1M

feat: MCP tools dynamic injection into agent TOOL_DEFINITIONS

Fix settings API: prevent broken apiKey/URL from causing 400, max tokens 1M, remove auto-compact, real Agent/PR/Issue, mobile fixes

Major update: fix 400 error, 1M context, real Agent/PR/Issue, no auto-compact

Fix 400 error (tool_calls JSON parse), rewrite WebFetch/WebSearch 1:1 from original claw-code

fix: mobile responsive, max tokens 128k, HuggingFace models, Quick Select UI

fix: add retry with exponential backoff for 429 errors, revert default to MiMo-V2-Flash

fix: chat auth bypass for dev mode, race conditions, model defaults

Claw Web v1.0 — AI Agent Web Interface with MiMo-V2-Flash