Menu
Show the user what mandelbot can do. Use this on first launch to introduce key features and workflows.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
| name | mandelbot-features |
| description | Show the user what mandelbot can do. Use this on first launch to introduce key features and workflows. |
| allowed-tools | [] |
Give the user a brief, friendly overview of what mandelbot can do. Cover these key points conversationally — don't just dump a feature list:
Mandelbot organizes work into a hierarchical tree of agents. The home agent (this tab) sits at the root. From here you can spawn project agents scoped to a working directory, and projects can spawn task agents for individual changes. Each agent has its own context, history, and working directory.
spawn_tab MCP tool to create new agents programmaticallycontrol+Space (default: ctrl+shift+Space) to spawn an agent interactivelycontrol+→ to spawn a child task under the current agentmovement+↑ / movement+↓ — move between sibling tabsmovement+← / movement+→ — move up/down the tree (parent / first child)movement+Space — jump to the next agent that needs attentionmovement+- — toggle back to the previously focused tabmovement+0-9 — jump to a tab by indexThe default modifier prefixes are ctrl+shift (control) and alt+shift (movement). These can be changed in ~/.mandelbot/config.json.
When a parent tab accumulates many children, use control+f to fold them into a single compact placeholder. This collapses the entire subtree so the tab bar stays manageable. Pressing any key while the fold placeholder is focused will unfold the tabs again.
In git projects, task agents automatically get their own worktree — an isolated copy of the repo where they can make changes without interfering with each other or the main working tree. Worktrees are created when a task spawns and cleaned up when the tab closes. The location is configurable via worktree_location in ~/.mandelbot/config.json (default: .mandelbot/worktrees relative to the project root).
When a task is large enough to parallelize, agents can use the /mandelbot-delegate skill to break work into subtasks and coordinate child agents via a shared status file. Each subtask gets its own worktree and branch.
Run /mandelbot-config to change theme, font, font size, shell, or keybinding prefixes. Settings live in ~/.mandelbot/config.json.
Not everything needs an agent. Use control+t to open a plain shell tab for quick terminal work.
Keep the overview concise and welcoming. The user just installed mandelbot — help them feel oriented without overwhelming them.
Use this skill from a project or task tab to map the user's open PRs onto a tree of tabs — one tab per conceptual group of authored PRs (with per-PR leaf tabs), plus a separate subtree for PRs awaiting the user's review. Good for starting the day, triaging in-flight work, or catching up after being away.
Use this skill when you have work that can be broken into parallel subtasks and delegated to child agents. Activates when you need to coordinate multiple agents working on different parts of a plan simultaneously.
Use this skill when your prompt references a coordination file and assigns you a task. You are a subtask agent — part of a larger coordinated effort. Follow the protocol to read your assignment, do your work, and report progress.
Use this skill from a project tab when the user wants to be notified about GitHub PR activity in this project's repo — a PR newly requests their review, or a PR they authored gets a review decision change. Polls `gh` in the background and raises one toast per affected PR via the `notify` MCP tool; the Open button spawns a task-tab child in this project's worktree.
Use this skill when correctness is adversarial — a persistent builder agent writes code while fresh breaker agents each round probe it and report what's broken, looping until the breaker reports clean or a round cap is hit. Good fit for algorithm-heavy, parser, or security-adjacent work where "what could go wrong" is the real risk and a single implementer is likely to miss cases.
Use this skill for iterative build-refactor-build loops where a parent agent delegates work to generation tabs, harvests ideas, and spawns another generation acting on the best ones. Activates when the user wants to "keep improving X", run "another pass", do "iterative refinement", or loop until some fitness signal is met. Not for one-shot feature work — use mandelbot-delegate for that.