Transforms analyzed data into polished, structured Markdown reports. Activates when the user wants to write a report, draft sections, create a narrative from data, or asks 'turn this analysis into a report.' Covers report structure, tone calibration, data narration, heading hierarchy, and audience-appropriate length.
From founder-osnpx claudepluginhub thecloudtips/founder-os --plugin founder-osThis skill uses the workspace's default tool permissions.
Designs and optimizes AI agent action spaces, tool definitions, observation formats, error recovery, and context for higher task completion rates.
Enables AI agents to execute x402 payments with per-task budgets, spending controls, and non-custodial wallets via MCP tools. Use when agents pay for APIs, services, or other agents.
Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.
Transform the structured analysis output from the Analysis Agent (step 2) into a complete, polished Markdown report ready for the Formatting Agent (step 4) to enhance with charts and visual elements. Treat every report as a standalone document a reader can understand without access to the underlying data. Lead every section with its most important finding. Support every claim with specific numbers from the analysis. Maintain a consistent professional tone throughout while adapting formality to the specified audience. Produce the full Markdown document in a single pass -- never leave placeholder sections or TODO markers. When the analysis output is missing data for a planned section, either omit the section or include a brief note explaining the data gap. Never fabricate data or invent statistics to fill gaps.
Organize every report into the following sections in this exact order. Omit a section only when the analysis output provides insufficient data, and note the omission in the report metadata.
Place first, always. Limit to 1-2 paragraphs (200-400 words). Write it last to ensure it reflects the full report. Structure in three beats: (1) one sentence of context stating purpose and scope, (2) two to four sentences covering the most significant findings, (3) one to two sentences identifying the highest-priority recommended actions. Never introduce information absent from the body.
Establish context in 150-300 words covering three elements: Context (why this report exists -- business question, milestone, or reporting cadence), Objectives (specific questions the report answers, framed as hypotheses when possible), and Scope (time period, data sources, entities covered, and explicit exclusions). Avoid previewing findings.
Describe data collection and analysis in 200-400 words. Include data sources by name (CSV files, Notion databases, Drive documents), a concise summary of each analytical technique the Analysis Agent applied, the exact date range, and known limitations or data quality caveats placed at the end of the section.
Present results organized by theme or business question -- never by data source. Open each subsection with the most important result as a topic sentence. Group related metrics under shared headings ("Revenue Performance" not separate sections per metric). Present quantitative findings before qualitative observations. Use tables for dense comparisons and inline bold for critical numbers. Order subsections by business impact. Limit the body to the top 8-12 results; relegate secondary findings to Appendices.
Interpret findings by answering three questions: What do they mean? (business implications tied to strategic goals), How do they compare? (vs. previous period, target, industry benchmark -- always state the baseline explicitly), and What caused them? (evidence-based explanations, distinguishing confirmed causes from hypotheses using "likely driven by" or "correlates with" for unconfirmed links). Keep discussion at 60-80% of Findings length. Reference numbers briefly rather than repeating them.
Structure each recommendation with four components: Action (clear imperative), Rationale (one sentence linking to a finding), Priority (High / Medium / Low), and Expected Impact (quantified when data supports it). Order by priority. Limit to 3-7 in the body; overflow to Appendices. Never recommend actions lacking supporting evidence.
Include raw data tables, a glossary of terms and acronyms, extended methodology details, secondary findings, and a complete data source inventory. Label each appendix with a letter (Appendix A, B, etc.) and reference from the main body.
Never present raw numbers without narrative context. Transform every data point into a statement that tells the reader what the number means for the business.
Pair every metric with at least one comparison point. Acceptable comparison baselines include:
When multiple comparisons are available, lead with the most relevant one for the audience and include others in supporting sentences.
Supplement absolutes with relative framing: "2x the previous quarter," "nearly half the industry average," "the largest single-quarter increase in two years." Especially valuable when readers lack baseline familiarity with the metric.
Flag data points deviating significantly from the trend. State the anomaly, quantify the deviation, and provide an explanation or explicitly note its absence: "February saw a 40% spike in support tickets (1,240 vs. the 6-month average of 885); the cause warrants further investigation."
#): Report title only. One per document.##): Major sections (Executive Summary, Introduction, Methodology, Findings, Discussion, Recommendations, Appendices).###): Subsections within major sections.####): Use rarely; prefer restructuring or bold lead-ins over deep nesting.Never skip heading levels. Keep heading text under 8 words. Use noun phrases or findings as headings, not questions or vague labels.
Total range: 1,800 words (single-theme) to 4,500 words (complex multi-theme). When exceeding 4,500 words, split into primary and supplementary reports and flag the decision in metadata for the QA Agent.
Produce a complete Markdown document using standard syntax: # headings, pipe tables with header separators, **bold** for critical numbers, bullet lists for enumerations, numbered lists for prioritized items. Do not include Mermaid charts, styled HTML, or complex formatting -- the Formatting Agent handles that. Leave chart placement markers using <!-- CHART: [description, data series, chart type] -->.
Include a metadata block at the end of the document:
Report Metadata:
Generated: [timestamp]
Data Sources: [count and names]
Analysis Themes: [count]
Total Findings: [count]
Recommendations: [count]
Word Count: [approximate]
Sections Omitted: [list any omitted sections and reason]
When the analysis output contains findings that contradict each other (e.g., revenue up but profit down, or two data sources reporting different values for the same metric):
Collapse Findings and Discussion into "Findings and Analysis." Maintain all other sections. Target 1,800-2,500 words total.
When analysis produces more than 15 findings, include the top 8-10 in the body ordered by impact. Move the rest to an "Additional Findings" appendix. Note the prioritization in the Executive Summary.