Agents
Agents d’IA i automatització
find-skills
iannuttall
Find and install agent skills with npx playbooks find skill and npx playbooks add skill. Use whenever a skill needs to be discovered or installed.
autonomous-execution-mandate
IgorGanapolsky
Self-directed autonomous execution mandate and workflow skill for ThumbGate agents. Enforces zero-human-intervention task chaining, proactive CI/Trunk monitoring, automated conflict resolution, and evidence-first reporting without requiring manual prompt nudges from the user.
context-engineering-checklist
IgorGanapolsky
HARD checklist before new skills, MCP tools, hooks/gates, plugins, or multi-agent workflows. Maps Hugging Face Context Course units 1–5 onto ThumbGate. Auto-invoke when creating SKILL.md, PreToolUse hooks, gate patterns, MCP tools, Grok workflows, slash commands, or when user says context engineering, HF context course, GSD, Ralph Loop, or agent context. Slash: /context-engineering-checklist.
withInterview
jthorare
"Interview the user to clarify requirements before executing a prompt. Use when the user wants to refine a task through guided questions before implementation."
thumbgate
IgorGanapolsky
Stop your AI from making the same mistake twice. Pre-Action Gates that block repeat hallucinations, retry loops, and known-bad tool calls before they reach the model — zero tokens spent on mistakes you've already corrected. Works with Claude Code, Cursor, Codex, Gemini CLI, Amp, OpenCode, and any MCP-compatible agent.
enhancePrompt
jthorare
Use when the user invokes /enhancePrompt — discovers available project documentation, reads relevant files, and rewrites the user's request with richer context before execution
thumbgate-feedback
IgorGanapolsky
Capture thumbs up/down feedback into structured memories and prevention rules. Use when user gives explicit quality signals about agent work (e.g. "that worked", "that failed", "thumbs up/down"). Do NOT use for general questions, code generation, file operations, or any task that is not explicit feedback on prior agent output.
jit-harness-compare-not-clone
IgorGanapolsky
JIT-Agent (arXiv:2608.25593) is a harness-intelligence model, not a ThumbGate clone. Steal the four-module FORMAT (memory/planning/action/capability) onto existing rails; never train or download JIT-Agent. Slash: /jit-harness-compare-not-clone.
thumbgate
IgorGanapolsky
ThumbGate provides pre-action gates for AI coding agents. It captures thumbs-up/down feedback on agent actions, auto-promotes repeated failures into prevention rules, and blocks known-bad tool calls via PreToolUse hooks. Trigger when the user wants to add safety guardrails to an AI agent workflow, capture structured feedback on agent output, generate prevention rules from failure patterns, gate high-risk actions before execution, or export DPO training pairs from production feedback. Works with any MCP-compatible agent including Cursor, Codex, Gemini CLI, Amp, and OpenCode.
Agent Memory
IgorGanapolsky
Give your agent persistent cross-session memory — recall what went wrong on similar tasks before starting, then capture whether the work succeeded or failed so repeated mistakes auto-promote into prevention rules. Runs locally via the ThumbGate MCP server (recall, capture_feedback, prevention_rules, feedback_stats) — no API key needed. Use when the user says "recall", "what went wrong last time", "remember this for next time", "capture feedback", "thumbs up/down on that", or at the start of a new task or session. Do NOT use when the user wants to force one specific mistake into an enforced block gate (use the thumbgate-guard skill), to list rules that already exist (thumbgate-rules), or to check ThumbGate wiring (thumbgate-doctor).
programmatic-agent-runs
IgorGanapolsky
Govern Cursor SDK local, cloud, self-hosted, and subagent coding runs before they create branches or PRs.
skill-forge
frankxai
Forge a reusable, portable skill from a proven workflow or capability brief. Use when the work needs repeatable instructions, triggers, guardrails, and evidence tests but does not need a persistent autonomous identity.
geo
yaojingang
Route GEO and generative engine optimization requests to an available GEOHub capability. Use for broad GEO requests, GEOHub workflow selection, capability checks, or requests spanning discovery, brand/site/page diagnosis, multi-page website diagnosis, content, strategy, knowledge, publishing, and measurement.
creative-director-draft
frankxai
Compile, inspect and revise the Starlight creative-director draft example without provider execution. Use for typed world or creator workflow planning, dependency invalidation, context selection and portable draft exports in this directory.
orchestration/multi-agent-coordination
frankxai
Use when multiple Starlight agents need to collaborate on a related task — handles agent selection, task decomposition, dependency management, and result aggregation across the council. Default skill for Starlight Orchestrator.
workflow-skill-creator
frankxai
Compatibility alias for the deprecated /workflow-skill-creator workflow. Route new requests through foundry/skill-forge for typed contracts, portable packaging, and receipt-based proof.
debate-coach
MoonTzai
《辩论筑基》体系辩论审问——引导出论与攻防处理。基于 grill-me 审问模式，完整内置《辩论筑基》(Debate Universal Grammar) 全套体系。自包含，不需另行学习。
intelligence/hermes-search
frankxai
Use when you need to find, retrieve, or research anything across vaults, repos, or the web — executes semantic search with full provenance, deduplicates across sources, flags contradictions, and returns ranked results sourced to the exact vault atom, file, or URL. Default skill for Starlight Hermes; callable by all agents.
orchestration/hermes-swarm
frankxai
Use when dispatching the Hermes agent swarm for any multi-IS or complex task — routes to IS-specialist Hermes profiles, coordinates parallel execution, aggregates synthesis, and writes results to vaults. Default skill for Starlight Hermes, Starlight Orchestrator, and Hermes Claw.
Check Agent Performance
OmniNode-ai
Agent routing and execution performance metrics including routing times, confidence scores, and transformation success rates
/onex:self_healing_dispatch — Self-Healing Orchestration Wrapper
OmniNode-ai
Self-healing orchestration wrapper — accepts ticket list or epic ID, groups by repo, dispatches via TeamCreate (enforced), monitors workers via agent_healthcheck, auto-recovers stalls with bounded retry (max 2 redispatches per task), logs all events to structured NDJSON.
Rewind — Conversational Rewind
OmniNode-ai
Query what a named agent was doing at a specific time. Conversational rewind — search the event stream by agent identity and timestamp.
agent-creator
frankxai
Compatibility alias for the deprecated /agent-creator workflow. Route new agent requests through foundry/agent-forge so persistent autonomy must be justified and independently proven.
creating-agent-teams
ZoranSpirkovski
Use when facing a complex task that could benefit from multiple agents, parallel work, or specialized roles - triggers on "create a team", "parallelize", "multi-agent", or cross-layer implementation needs