The daily firehose — everything the toolchain shipped today, already filtered.
// HOW THIS ISSUE IS MADE
We read every release from the 174 tools on our watchlist at the source — GitHub and GitLab release notes, vendor release pages and changelogs, project blogs and feeds, vendor press releases, and the source code behind the tag. Bug-fix-only releases and non-product newsroom noise are dropped; what's left is summarized down to the new capability, how to try it, and any screenshots or videos the release itself published. Every entry links to the sources it was built from.
Hermes v2026.3.30 adds multi-instance profiles, MCP server mode, Feishu/WeCom/Slack multi-workspace support, fallback provider chains, and Exa search.
└──▷ GET THIS VERSION
$ git clone --branch v2026.3.30 https://github.com/NousResearch/hermes-agent.git
# already have the repo? check out this version:$ git checkout v2026.3.30
└──▷ TRY IT
Run isolated staging and production agent instances from a single Hermes installation without credential collisions.
$ hermes profile create staging
hermes -p staging model set nous/hermes-3-70b
hermes -p staging chat
Keep agent tasks running even when your primary provider is rate-limited or down by configuring an automatic failover chain.
›New hermes profile create/list/switch/delete/export/import/rename commands enable fully isolated multi-instance deployments, each with its own config, memory, sessions, skills, and gateway service with token-lock credential isolation.
›New hermes mcp serve exposes conversations, sessions, messages, and attachments to any MCP-compatible client (Claude Desktop, Cursor, VS Code) via stdio or Streamable HTTP transports.
›Official Dockerfile enables running Hermes Agent in a container in both CLI and gateway modes with volume-mounted config.
›New fallback_providers config key enables an ordered failover chain across multiple inference providers.
›Adds full gateway adapter for Feishu/Lark with event subscriptions, message cards, group chat, image/file attachments, and interactive card callbacks.
+12 moreshow less
›Adds new gateway adapter for WeCom (Enterprise WeChat) supporting text, image, and voice messages, group chats, and callback verification.
›Slack gateway now supports multi-workspace OAuth, connecting a single gateway to multiple workspaces via an OAuth token file with per-workspace bot token resolution.
›Telegram adapter gains webhook mode as a production alternative to polling, plus configurable group mention gating (always, @mention-only, or regex-triggered).
›Adds Exa as an alternative web search and content extraction backend alongside Firecrawl and DuckDuckGo, configured via EXA_API_KEY.
›Skills and credential files can now be mounted into Modal and Docker remote backends so remote terminal sessions share the same skills and secrets as local execution.
›Discord gateway adds a DISCORD_IGNORE_NO_MENTION env variable to skip messages that @mention other users or bots but not Hermes.
›Discord gateway adds a processing reaction emoji that appears while a request is in flight and is removed on completion.
›Matrix adapter sends voice messages as native Matrix voice events via MSC3245 instead of file attachments.
›Mattermost adapter gains configurable mention behavior to respond without requiring an @mention.
›Gemini 3.1 preview models added to OpenRouter and Nous Portal catalogs.
›Rate limit (429) responses now display a user-friendly message with a Retry-After countdown.
›Curated model lists are now preferred over live API probes when the probe returns fewer models.
└──▷ BREAKING ON UPGRADE
!When no provider is configured, Hermes now raises a clear error instead of silently routing to OpenRouter.
Copilot CLI v1.0.13 adds conversation rewind timeline, MCP LLM sampling requests, and faster startup via V8 compile cache.
└──▷ GET THIS VERSION
$ git clone --branch v1.0.13 https://github.com/github/copilot-cli.git
# already have the repo? check out this version:$ git checkout v1.0.13
└──▷ TRY IT
Roll back a conversation to an earlier point when an agent went down a wrong path.
$ /rewind
›New /rewind command and double-Esc gesture open a timeline picker that can roll back to any point in conversation history, not just the previous snapshot.
›MCP servers can now request LLM inference (sampling) with user approval via a new review prompt.
›MCP servers blocked by allowlist policy are now hidden from /mcp show.
›CLI startup time reduced by running terminal detection, auth, and git operations in parallel, and by a V8 compile cache that cuts parse and compile time on repeated invocations.
›MCP server OAuth authentication now works when the CLI runs in ACP mode.
+3 moreshow less
›MCP registry lookups gain automatic retries and request timeouts for improved reliability.
›Reasoning effort setting now applies correctly when using Bring Your Own Model (BYOM) providers.
›Splits $BROWSER on spaces to support browser launch configurations that include arguments.
└──▷ BREAKING ON UPGRADE
!Support for the gemini-3-pro-preview model has been removed.
OpenHands 1.6.0 adds hooks support, /clear and /new chat commands, toggleable global skills, and Catalan UI localization.
└──▷ GET THIS VERSION
$ git clone --branch 1.6.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:$ git checkout 1.6.0
└──▷ HOW TO FIND IT
Reset chat context mid-session without losing your running sandbox state — useful when a conversation has drifted and you want a clean prompt history.
📍In the chat input, type /clear and press Enter. A new conversation opens, inheriting the current sandbox and configuration.
›Adds hooks support, enabling extensible event-driven integrations with the OpenHands runtime.
›New /clear slash command starts a fresh conversation that inherits the current sandbox and configuration, preserving runtime state while resetting chat history.
›New /new slash command is now surfaced in the slash command menu for faster access to new conversations.
›Adds ability to enable or disable default global skills, giving operators control over which built-in capabilities are active.
›Adds a copy button to code blocks in the chat UI for easier extraction of generated code.
Qwen Code SDK TypeScript v0.1.6 adds session resumption, multi-modal input, concurrent batch runner, LSP support, and a new settings.env field.
└──▷ GET THIS VERSION
$ git clone --branch sdk-typescript-v0.1.6 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:$ git checkout sdk-typescript-v0.1.6
└──▷ TRY IT
Resume a previous session by ID to continue a long-running task without losing context.
$ qwen --session-id <your-session-id>
Configure environment variables for the tool directly in settings, avoiding shell profile changes.
›Adds --session-id CLI flag and resume/continue SDK options to resume or continue existing sessions by ID.
›Adds settings.env field in settings configuration for environment variable configuration.
›Adds extra_body support for OpenAI-compatible providers to pass additional request body parameters.
›Adds contextWindowSize configuration key for controlling model context window size.
›Adds export command for session history, supporting markdown and HTML output formats.
+22 moreshow less
›Adds parentToolCallId and subagentType fields for ACP subagent tracking.
›Adds FORK_MODE support to ProcessTransport for Electron IPC integration.
›Adds symlink support for the skill manager.
›Adds MCP tool progress update support in TUI and SDK mode.
›Adds cache_control for system and last user text messages in the Anthropic content generator to optimize prompt caching.
›Adds Coding Plan authentication mode with a unified AuthDialog and configuration management.
›Adds TPM throttling error handling with a 1-minute retry delay.
›Adds large paste placeholder and improved enter-submit behavior on macOS.
›Adds a concurrent runner for batch CLI execution.
›Adds experimental LSP support for code intelligence.
›Adds multi-modal input support (image, PDF, audio) across all content generators.
›Adds Zed extension for the Qwen Code agent server.
›Promotes Agent Skills from experimental to stable.
›Adds support for model selection through ACP in the VS Code IDE companion.
›Preserves UTF-8 BOM when editing files.
›Improves retry logic for 429/5xx error handling.
›Removes the Smart Edit tool and ClearcutLogger.
›Removes the read_many_files tool and replaces it with a readManyFiles utility for user @-commands.
›Routes debug mode console calls to a logfile-first debugLogger.
›Renames negative settings from disable* to enable* naming convention.
›Clarifies output formats for non-interactive mode.
›Adds source information tracking in telemetry logs.
└──▷ BREAKING ON UPGRADE
!Settings keys previously named with a disable* prefix are renamed to enable* — any existing configuration using the old disable* setting names will need to be updated.
!The read_many_files tool is removed; functionality is replaced by a readManyFiles utility available via user @-commands.
!The Smart Edit tool is removed; workflows relying on it will need to use alternative edit tools.
llama.cpp server gains header wrapping for the MCP CORS proxy
└──▷ GET THIS VERSION
$ git clone --branch b8581 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:$ git checkout b8581
›Adds header wrapping in server-cors-proxy.h for the MCP proxy, enabling proper CORS header forwarding when the llama.cpp server is used as an MCP proxy.
Phoenix 13.20.0 adds REST endpoints to delete prompts and prompt version tags, plus a new copy field in the settings UI.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v13.20.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-v13.20.0
└──▷ TRY IT
Remove a tag (e.g. 'production') from a prompt version to unpin it from that lifecycle stage.