Loading...
Loading...
Found 13,353 Skills
Update an existing AI agent on a live LiteLLM proxy. Ask for the agent_id and what to change (model, description, MCP servers), then call PATCH /v1/agents/{agent_id}.
Use when adding, pruning, archiving, restoring, or reviewing Agent Notes — checking new notes for superseded active records, classifying implemented notes by future decision value, deleting rejected notes that no longer prevent a tempting mistake, and moving low-value notes to the frozen archive.
Ask which skill or flow fits your situation. A router over the docs/Agent-Note maintenance flow, the documentation governance run, the prose-and-docs tool layer, and the standalone git and game skills in this pack.
Configure a repository for the docs-maintenance skills — scaffold the Agent Note decision tree, the bilingual pairing convention, the documentation standard, the session-discipline lines, the documentation budget gate, the change-scope tool, and the git hooks, and record where they live. Run once before first use of the maintenance flow.
Use when working in a repository to find non-obvious simplification candidates — dead, duplicated, speculative, over-built, added-then-removed, or hand-rolled-where-a-dependency-exists surface across code, docs, and skills — and turn the worthwhile ones into proposed Agent Notes or inline TODO/FIXME/XXX markers; also when asked to audit or coalesce superseded Agent Notes, or fold worthwhile simplification ideas from another PR or branch.
Use when the user is leaving and wants work carried to completion without them — "I'm going to sleep", "work overnight", "keep working while I'm away", "I'll be offline", "don't wait for me", "don't ask, just do it", "merge it yourself", "full autonomy". Also use when a long task must survive context compaction or session death with an audit trail.
Generate targeted test scenarios for a LiveKit voice or chat agent and run them as simulations — locally, from the agent's own code plus what the user wants stress-tested. Use whenever the user wants to "test my agent", "what should I test", "create/generate simulation scenarios", "make a sim test suite", "use lk agent simulate", "stress-test the X flow", "set up scenarios for my agent", or wants to probe edge cases / refusals / regressions before shipping. Generates scenarios on the user's machine (their code is never uploaded) and lets the user deeply steer what gets tested. Trigger even without the word "simulation" when the user clearly wants to decide what to test and verify how their agent behaves across realistic conversations. Not for building a new agent from scratch (use the livekit-agents skill), load-testing, or ordinary unit tests.
Review a pull request or a branch against the repository's standards — skill frontmatter and body, docs and bilingual pairs, Agent Notes, and scripts — and report blockers separately from suggestions, prioritizing correctness and required behavior over style.
Use when auditing or fixing prose that reads like a leaked reasoning transcript — dead session citations, change narration, review residue, or hedged planning in docs, skill bodies, comments, or Agent Notes.
Install, update, or uninstall hook-based desktop/audio reminders for Agent stop and attention events. Use only when the user explicitly invokes agent-doorbell or asks to configure, enable, install, update, disable, remove, or uninstall automatic Agent doorbell hooks. Do not use for ordinary final responses or manual stop ringing.
Use when writing or reviewing agent instructions and current-state docs — rules files (AGENTS.md/CLAUDE.md), skill bodies and descriptions, prompts, and design/plan/process docs — or on request to prune a task or session file. Restates negation priming, absence declarations, stale meta-narrative, and strawman warnings as positive targets or consequences.
The canonical way to run agent-relay - self-bootstrap the local broker and autonomously spawn, monitor, and coordinate a team of worker agents without human intervention. Covers infrastructure startup, agent spawning, lifecycle monitoring, message-based reading via the relay MCP, and team coordination.