# Structure Causal Business Decisions

Business teams often mix facts, assumptions, decisions, and outcomes in one discussion. This skill separates each step into a traceable causal workflow that supports better execution and learning.

## Install

```bash
npx skillstore add bellabe/reasoning-causal
```

## Metadata

- - Slug: bellabe-reasoning-causal
- - Version: 1.0.0
- - Author: BellaBe
- - GitHub username: BellaBe
- - License: MIT
- - Repository: https://github.com/BellaBe/lean-os/tree/main/.claude/skills/reasoning-causal
- - Ref: main
- - Supported tools: Claude, Codex, Claude Code
- - Risk level: medium
- - Risk factors: filesystem
- - Quality score: 50
- - Quality tier: warning
- - Public page: https://skillstore.pages.dev/skills/bellabe-reasoning-causal
- - Manifest: https://skillstore.pages.dev/api/skills/bellabe-reasoning-causal/manifest

## Capabilities

- Defines a six-stage flow from input capture through learning updates.
- Provides templates for observations, hypotheses, implications, decisions, actions, and outcomes.
- Links decisions to business assumptions, impact scores, goals, and Canvas updates.
- Separates operational thread types for sales, marketing, engineering, and business decisions.
- Includes guidance for human review when impact or risk is high.

## Use Cases

- Validate a Product Assumption: Turn customer feedback into a hypothesis, quantify impact, and decide whether to build or defer.
- Coordinate Operational Execution: Break an approved decision into assigned actions with owners, deadlines, dependencies, and success criteria.
- Close the Learning Loop: Compare expected and actual outcomes, update assumptions, and record follow-up opportunities.

## Prompt Templates

### Capture an Observation

```
Use reasoning-causal to create Stage 1 for this observation: [paste fact]. Keep interpretation and recommendations out of the input.
```

### Test a Business Assumption

```
Use reasoning-causal Stage 2 to map this observation to our assumptions. Identify what is validated, challenged, or newly hypothesized.
```

### Quantify Decision Impact

```
Use reasoning-causal Stages 3 and 4 for this proposed decision. Include alternatives, ROI, approval threshold, and human review status.
```

### Run a Full Decision Thread

```
Use reasoning-causal to build a full thread for [initiative]. Create stages for input, hypothesis, implication, decision, actions, and learning.
```

## Limitations

- It is a Markdown guidance skill and does not include executable automation.
- Users must supply accurate business data, evidence links, costs, and outcomes.
- High-impact decisions still need human review and organizational approval.
- The workflow assumes a Canvas and goal-tracking structure exists in the workspace.

## Best Practices

- Start with evidence and dates before asking for conclusions or recommendations.
- Require human approval for decisions with high cost, high risk, or strategic impact.
- Review generated file changes before committing updates to strategy or goal documents.

## Anti Patterns

- Do not skip from customer feedback directly to action planning.
- Do not treat estimates as facts without naming assumptions and confidence levels.
- Do not allow autonomous execution for commitments that need legal, financial, or executive approval.

## Security Audit

- - Safe to publish: true
- - Audited at: 2026-06-28T15:06:03.077\+00:00
- - Summary: Static analysis reported many external command and weak cryptography patterns, but sampled evidence shows Markdown code fences, inline paths, and business templates rather than executable code. No confirmed malware, prompt injection, network exfiltration, or real shell execution was found. The remaining risk is that the skill instructs agents to create and update operational files and may allow autonomous business decisions, so publication should include a filesystem and human-review warning.

## Stats

- - Views: 489
- - Downloads: 5
- - Favorites: 0
- - Popularity score: 0