4,052 tools found
Makes n8n failures loud and recoverable - per-node error outputs, retryOnFail self-healing, status-coded responses, and a workflow-level error catch-all.
Prevents silent misroutes when one MCP session targets multiple n8n instances - discover, switch, and verify before every write.
Design n8n sub-workflows as typed, discoverable functions: contract-first inputs/outputs, correct call mode, and agent-tool wiring.
Router skill that names which n8n-mcp specialist skill to invoke first, before any workflow design, edit, validation, or credential action.
Rigorously test if a cheaper LLM can replace your expensive pipeline with multi-seed sweeps, error bars, and honest statistical ties.
Generates decision docs, async standups, and status updates using remote-team best practices from the Open and Async playbook.
Self-hosted control plane and registry for internal AI Skills, MCP servers, and agents, with per-harness config generation and usage insights.
LoopGain is an open-source cost controller that stops AI agent loops the moment they converge, instead of running to a fixed max_iterations cap.
A skill that guards generated code against 24 clean-code and AI-specific failure imperatives before it ships.
A 6-phase, multi-agent audit methodology for authorized codebases: reconnaissance, adversarial hunting, validation, and reproducible findings only.
A skill that researches competitor Meta and Google ads and produces a strategic teardown with hooks, funnels, and counter-plays.
A skill for diagnosing Android device overheating via ADB - read-only baseline capture, controlled comparisons, and safety gating.
