FPEF Evidence Analyzer

Files (3)

• SKILL.md
• fpef-config.json · resources/
• fpef-analyze.sh · scripts/

SKILL.md (first lines)

# FPEF Evidence Analyzer

## Overview
Rigorous evidence-based analysis framework that systematically finds problems, proves root causes, gathers supporting evidence, and implements targeted fixes for complex technical and business systems.

## Prerequisites
- System or problem description
- Access to relevant data sources (logs, metrics, code)
- Basic understanding of the domain being analyzed

## What This Skill Does
1. **Find**: Systematic identification of anomalies and potential issues
2. **Prove**: Causal analysis to establish root cause relationships
3. **Evidence**: Comprehensive evidence collection and validation
4. **Fix**: Targeted interventions with verification of effectiveness

---

## Quick Start (60 seconds)

### Rapid Analysis
```bash
# FPEF Interactive Analysis
1. Describe the problem: [e.g., "API response times increased 300%"]
2. Timeframe: [e.g., "Last 24 hours"]
3. Affected systems: [e.g., "Payment processing API"]
4. Available data: [e.g., "CloudWatch logs, database metrics"]

→ FPEF generates comprehensive analysis framework
```

### Immediate Output
Analysis framework includes:
- ✅ Problem statement with clear scope definition
- ✅ Hypothesis tree with potential root causes
- ✅ Evidence collection plan with data sources
- ✅ Investigation timeline with critical path
- ✅ Fix validation strategy and success criteria
- ✅ Prevention measures for future incidents

---

## Configuration

### Analysis Parameters
Edit `resources/fpef-config.json`:
```json
{
  "problem_description": "Clear description of the issue",
  "scope": "Systems, components, or processes affected",
  "timeline": "When the problem started and duration",
  "severity": "critical|high|medium|low",
  "business_impact": "Revenue, users, operations affected",
  "data_sources": ["logs", "metrics", "traces", "code"],
  "constraints": ["time", "budget", "access", "expertise"]
}
```

### Analysis Settings
```json
{
  "depth": "comprehensive|focused|rapid",
  "certainty_threshold": "0.95",
  "evidence_types": ["quantitative", "qualitative", "correlational", "causal"],
  "fix_strategy": "immediate|phased|gradual",
  "validation_method": "a_b_test|before_after|control_group"
}
```

---

## Step-by-Step Guide

### Phase 1: FIND - Systematic Problem Identification (5 minutes)

#### Step 1.1: Problem Statement Definition
FPEF structures clear problem statements:
- **What**: Specific deviation from expected behavior
- **Where**: Systems, components, or processes affected
- **When**: Timeline and frequency of occurrence
- **Impact**: Business and technical consequences
- **Metrics**: Quantifiable measures of the problem

#### Step 1.2: Scope Analysis
```bash
# Generate scope map
./scripts/fpef-find.sh --scope-analysis

# Outputs:
# - System boundary definition
# - Stakeholder impact matrix
# - Risk assessment and prioritization
# - Resource requirements for investigation
```

#### Step 1.3: Hypothesis Generation
FPEF creates structured hypothesis trees:
- **Primary Hypotheses**: Most likely root causes
- **Secondary Hypotheses**: Alternative explanations
- **Contributing Factors**: Multi-causal relationships
- **External Factors**: Environmental influences

### Phase 2: PROVE - Causal Analysis and Root Cause Identification (10 minutes)

#### Step 2.1: Evidence Planning
```bash
# Generate evidence collection plan
./scripts/fpef-prove.sh --evidence-plan

# Evidence categories:
# - Direct Evidence: Direct measurements of the problem
# - Correlational Evidence: Statistical relationships
# - Circumstantial Evidence: Contextual factors
# - Expert Evidence: Domain knowledge and experience
```

#### Step 2.2: Causal Chain Analysis
FPEF establishes causal relationships:
- **Temporal Sequence**: Verify cause precedes effect

From aegntic/cldcde, MIT.

More from aegntic/cldcde