Skills manager, per-server MCP controls, and status-pill transcript polish - #6
Merged
Merged
Conversation
…ranscript polish Skills - New "Skills" rail tab (graduation-cap icon) to manage global skills (~/.mandocode/skills), the reusable instruction sets the AI loads on demand. - Create / edit / delete skills in-app, with an optional "Draft with AI" panel that generates a skill's name, description, and instructions from a one-line intent using a user-picked model (defaults to the active agent's). - Install skills from a git URL, a .zip, or a local folder, with recursive discovery so nested repo layouts (e.g. skills/<name>/SKILL.md) work. - Enable/disable each skill (renames SKILL.md so the engine skips it) — changes fan out to every open agent immediately. - Search box, quick-filter chips (All/Enabled/Disabled/Large), grouped Enabled/Disabled sections, and a per-skill size estimate so heavyweight skills are easy to spot. MCP servers - Replaced the per-agent "Use MCP tools" opt-in with per-server enable/disable, since servers are shared by every agent. Toggling a server saves and restarts/reconnects it across all agents. - Same layout as Skills: search, filter chips (All/Enabled/Disabled/Failed), grouped sections, consolidated [actions | filters] toolbar, and a footer. Transcript polish - Tool calls render as a single static pill (shows the loaded skill's name); no animation, to avoid a prior CPU-spin regression. - Connection/status/confirmation output (MCP reload, /copy, /copy-code, /config, /music) render as centered, theme-aware status pills — no emoji. Lists, errors, and detail-bearing messages stay as readable text. All changes are confined to the Desktop app; the shared engine submodule is untouched.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a first-class way to manage the AI's skills and MCP servers from the desktop app, and cleans up how the transcript reports background/system activity. All work is in the Desktop app — the shared engine is not touched.
A "skill" is a reusable set of instructions the AI pulls in automatically when a request matches it; an "MCP server" connects the AI to outside systems (databases, APIs, docs) as tools it can call. Both are shared across every chat agent. Previously neither could be managed from the UI beyond a basic list.
What's new
Skills management (new "Skills" tab)
.zip, or a local folder (handles nested layouts).MCP server management
Transcript polish
User-facing behavior change to note
Scope & risk
How to review
Launch the app, open the Skills and MCP servers tabs from the left rail, and exercise create / enable-disable / search / filter / install, then run a chat that loads a skill and toggles a server to confirm the transcript pills and reload behavior.