Skip to content
Conduit

skill

problem-solving-systems-thinking

This skill helps an agent decompose complex problems, map workflows, identify decision points, and design an agent's end-to-end flow from intake through validation and final response.

Set up

Skills are installed by placing their folder where your agent loads skills.

The download comes from github.com/bybaochaule/problem-solving-systems-thinking itself, not from our server, so what you get is exactly what its author published.

Install instructions come from the README

Skills do not use an MCP config. Follow the steps in the README below.

README

synced Oct 22, 2026

Problem Solving & Systems Thinking

Overview

This skill helps an agent decompose complex problems, map workflows, identify decision points, and design an agent's end-to-end flow from intake through validation and final response.

Use it when a user asks for process design, systems thinking, workflow mapping, lead generation flow design, decision trees, SOP design, bottleneck analysis, or agent behavior design.

What it creates

The skill can produce:

  • Problem decompositions and system boundaries.
  • Current-state and future-state workflow maps.
  • Decision tables with rules, branches, thresholds, and escalation paths.
  • Agent flow specs covering how the agent thinks, decides, acts, validates, and escalates.
  • Mermaid diagrams, SOPs, state machines, RACI-style handoff tables, and implementation-ready process specs.
  • Risk, bottleneck, and missing-information analysis.

Example user prompts

  • "Map our lead generation process from prospecting to sales handoff."
  • "Decompose this messy customer onboarding problem and design the workflow."
  • "Create a decision tree for support ticket triage."
  • "Design how an agent should think, decide, and act for invoice review."
  • "Turn this process into a state machine with escalation paths."

Contents

  • SKILL.md: agent-facing instructions, trigger description, boundaries, workflow, decision policy, output formats, quality checks, and safety rules.
  • README.md: human-readable overview and usage notes.
  • agents/openai.yaml: OpenAI metadata, default prompt, invocation policy, and file dependencies.
  • references/style-guide.md: style rules, output patterns, terminology, edge cases, and anti-patterns.
  • references/lead-generation-example.md: worked example for mapping a lead generation workflow.
  • assets/workflow-map-template.md: reusable workflow mapping template.
  • scripts/example_helper.py: deterministic helper that checks whether a Markdown workflow map includes key sections.
  • assets/.gitkeep: keeps the assets directory in version control.

Package structure

problem-solving-systems-thinking/
|-- SKILL.md
|-- README.md
|-- agents/
|   `-- openai.yaml
|-- assets/
|   |-- .gitkeep
|   `-- workflow-map-template.md
|-- references/
|   |-- style-guide.md
|   `-- lead-generation-example.md
`-- scripts/
    `-- example_helper.py

Suggested default output

# [Process or Problem Name]

## 1. Goal and success metric
## 2. System boundary and assumptions
## 3. Actors, tools, inputs, and outputs
## 4. Workflow map
## 5. Decision points
## 6. Agent flow
## 7. Risks, bottlenecks, and mitigations
## 8. Suggested next steps

Safety notes

The skill is for analysis, design, and planning. It should not execute workflows, contact people, scrape data, update systems, or make high-stakes determinations without explicit authorization, safe tooling, and human review where appropriate.

↑↓ navigate openesc close