AI Agent Workflow Collection

Create 1-3 short asciinema demo recordings (5-20 seconds each) that demonstrate completed work. Use only when explicitly instructed to create demos.

Use when you want to create a new web view for the user. Covers scaffolding a new FastAPI service (canonical path) and the escape hatch for wrapping a pre-existing third-party server, plus diagnostic references when things misbehave.

Convert triaged autoclaude GitHub issues into FIXMEs in the codebase. Use when you want to process issues that have been triaged by authorized users.

Create FIXME's in the codebase for each of the issues in the given input file.

Convert a file containing identified issues into GitHub issues. Use after running identify- commands to create corresponding GitHub issues.

Create a prompt from a short task description in a JSON…

Turn a process from the turn that just finished into a reusable skill. A skill captures a stable process -- SKILL.md prose describing the recipe, with scripts for deterministic steps and prose instructions for nondeterministic steps. Consider using after completing a task where a re-run with new inputs would follow a largely similar process. The process does not have to be the entire turn -- a sub-process (e.g. a data pipeline within a larger build) counts. Strong signal: you learned how to do something through research or debugging that is likely to be useful again.

Handle unexpected situations where things are not working as expected. Use when you encounter errors, confusing state, or behavior that contradicts your docs and prompts.

Use immediately when the user asks you to do something net-new -- a task you haven't done before, no existing skill applies, and getting it right will require nontrivial research, exploration, or experimentation. When doing this, give a very short confirmation message to the user's request, then load this immediately before responding further. Your confirmation message shouldn't mention loading the skill. Skip when an applicable single skill already exists or for pure dev/code-writing work.

Add, modify, or remove background services managed by the bootstrap service manager. Use this when you want to run a long-lived process alongside your main agent.

Resolve an agent name or description to an exact mngr agent name. Used by other skills that target agents.

Identify places in the $1 library where the docs and code…

Identify inconsistencies in the $1…

Identify outdated docstrings in the $1…

Identify divergences from the style guide in the $1…

Use whenever you want to use latchkey commands or interact with third-party or self-hosted services (Slack, Google Workspace, Dropbox, GitHub, Linear, Coolify...) using their HTTP APIs on the user's behalf.

Create a sub-agent to perform a larger task. Use when work is large enough to warrant a separate context, involves multi-file changes, or benefits from isolation.

Send a message to another mngr agent. Use when you need to communicate with a peer agent.

End-to-end dev workflow for the minds app stack -- first-time bring-up, every-startup vendor/mngr sync, and the iteration loop against a running Docker agent. Use this when starting or restarting the dev Electron app, or after changing any minds component (mngr, the workspace server, the FCT template).

Create a new PRIVATE GitHub repo that is a full-history copy of imbue-ai/forever-claude-template's current main branch, clone it to / (default $HOME/project), and push. Use when the user asks to "spin up a new forever-claude clone", "fork the forever-claude template as a private repo", "make me a new private copy of forever-claude-template", or similar.

Telegram-specific implementation for reading recent conversation history. Only use when the deployment is known to use telegram (e.g. invoked from within send-user-message after its telegram probe succeeds). Deployments with other channels have their own equivalents.

Cut a new "production" release of the minds app. Pushes a release branch in the mngr clone at ~/project/mindsprod, syncs vendor/mngr in ~/project/forever-claude-template to match, pushes the same-named branch there, and merges the release branch into FCT main. Use when the user asks to "release a new version of minds", "cut a minds release", "update the vendored mngr in forever-claude-template to track ", or anything of that shape.

Telegram-specific implementation for sending a message to the user. Usually invoke send-user-message instead -- it probes for the configured channel and dispatches here only when telegram is active. Only invoke this skill directly when you already know the deployment uses telegram (or a caller skill tells you to).

Send a message to the user through whatever channel the deployment has configured. Use whenever another skill tells you to "ask the user" or "tell the user" something. Dispatches to a specific channel (e.g. telegram) if available, otherwise falls back to inline output in the current chat.