Report Methodology Skill

This skill provides reasoning guidance for report generation. It defines the eight report types, their purposes, audiences, tone requirements, and agent recruitment patterns.

Report Types and Purpose

Executive Summary

Purpose: High-level overview for leadership and non-technical stakeholders Audience: Executives, product leaders, business stakeholders Length: 2-3 pages Tone: Business-focused, clear, concise, minimal technical jargon Agents: investigator, synthesizer (analyst skipped for efficiency)

Key themes: Overview, key findings (3-5 points), recommendations, next steps

Technical Deep Dive

Purpose: Detailed technical analysis for engineering teams Audience: Engineers, architects, technical leads Length: 5-10 pages Tone: Technical, thorough, analytical Agents: investigator, analyst, synthesizer

Key themes: Context, technical analysis, implementation details, trade-offs, recommendations

Competitive Analysis

Purpose: Market and competitor research Audience: Product managers, executives, strategy team Length: 3-5 pages Tone: Strategic, comparative, insight-driven Agents: investigator, analyst, synthesizer

Key themes: Market context, competitor comparison, strengths & gaps, strategic recommendations

Architecture Review

Purpose: System design evaluation and recommendations Audience: Architects, senior engineers, technical leads Length: 4-8 pages Tone: Architectural, evaluative, forward-looking Agents: investigator, analyst, synthesizer

Key themes: Current architecture, strengths, weaknesses, recommendations, migration path

Performance Analysis

Purpose: Performance metrics, bottlenecks, and optimization opportunities Audience: Engineers, SREs, technical leads Length: 3-6 pages Tone: Data-driven, metrics-focused, actionable Agents: investigator, analyst, synthesizer

Key themes: Metrics, bottlenecks, root causes, optimization opportunities, implementation plan

Incident Postmortem

Purpose: Post-incident analysis and learning Audience: Engineering teams, leadership, SREs Length: 3-5 pages Tone: Objective, blameless, learning-focused Agents: investigator, analyst, synthesizer

Key themes: Incident summary, timeline, root cause, impact assessment, action items

Quarterly Review

Purpose: Periodic progress assessment Audience: Teams, leadership, stakeholders Length: 4-6 pages Tone: Reflective, progress-focused, forward-looking Agents: investigator, synthesizer (analyst skipped for efficiency)

Key themes: Period summary, achievements, metrics, challenges, outlook

Feasibility Study

Purpose: Evaluation of new initiative viability Audience: Product managers, executives, engineering leads Length: 5-8 pages Tone: Evaluative, balanced, recommendation-driven Agents: investigator, analyst, synthesizer

Key themes: Proposal overview, technical feasibility, resource requirements, risk assessment, recommendation (go/no-go)

Category Taxonomy

Categories organize reports by domain. Each report has exactly one primary category:

Agent Recruitment Logic

Different report types use different agent combinations:

investigator + synthesizer only (skip analyst for efficiency):

• •executive-summary – straightforward assembly, minimal interpretation needed
• •quarterly-review – progress report, straightforward data collection

investigator + analyst + synthesizer (full pipeline):

• •technical-deep-dive – needs detailed interpretation
• •competitive-analysis – needs strategic interpretation
• •architecture-review – needs pattern analysis and evaluation
• •performance-analysis – needs metric interpretation and root cause analysis
• •incident-postmortem – needs root cause analysis and learning extraction
• •feasibility-study – needs risk/opportunity assessment

Confidence Level Guidance

Set confidence levels based on data availability and investigation completeness:

High Confidence:

• •Comprehensive data available, all relevant sources examined
• •Clear patterns identified, minimal assumptions
• •Findings validated across multiple sources

Medium Confidence:

• •Moderate data availability, some gaps in information
• •Patterns identified but with caveats, some assumptions necessary
• •Findings partially validated

Low Confidence:

• •Limited data available, significant information gaps
• •Tentative patterns or unclear findings
• •Many assumptions required, findings not fully validated

Tone and Style by Report Type

Adopt the appropriate tone for each report type:

Executive Summary: Business language, clear value statements, avoid technical jargon, focus on outcomes and decisions

Technical Deep Dive: Technical precision, detailed explanations, appropriate use of technical terms, focus on implementation

Competitive Analysis: Strategic perspective, comparative framing, market context, focus on positioning and differentiation

Architecture Review: Architectural thinking, system-level view, design patterns, focus on structure and evolution

Performance Analysis: Data-driven, metrics-focused, quantitative, focus on measurements and optimization

Incident Postmortem: Objective, blameless, timeline-based, focus on learning and prevention

Quarterly Review: Reflective, balanced (achievements + challenges), progress-oriented, focus on trajectory

Feasibility Study: Evaluative, risk-aware, recommendation-oriented, focus on viability and go/no-go decision

Content Depth by Audience

Match content depth to the intended audience:

Leadership (Executives, Product Managers):

• •Start with high-level summary, focus on business impact
• •Minimize technical details, emphasize actionable recommendations
• •Use analogies and examples to explain complex concepts

Technical Teams (Engineers, Architects):

• •Provide technical depth and specifics
• •Include implementation considerations, discuss trade-offs and alternatives
• •Reference specific systems, APIs, libraries

Cross-functional (Mixed Audiences):

• •Layer information (summary first, details follow)
• •Define technical terms when first used
• •Balance business context with technical reality, provide both strategic view and tactical details