Report & Information Design

Turn findings into a document a busy decision-maker can scan in two minutes and a practitioner can act on for a month, and make it render flawlessly.

Principles

1. Lead with the conclusion (BLUF). Open with the finding and the recommendation.
2. One idea per block.
3. Design for scanning (headings, short paragraphs, callouts, visuals).
4. Show, don't tell (charts, annotated blueprints over prose).
5. Cite inline + collect sources with reliability tiers and the gapless citation spec (references/citations-and-grounding.md).
6. Prioritize visibly (evidence-strength × effort matrix + phased roadmap).
7. Calibrate to audience (internal-technical vs client-facing).
8. Self-contained, responsive, print-friendly, QA'd before delivery.

Default report structure

Executive summary → data/baseline → diagnosis (a model, not symptoms) → current-state teardown → evidence (tiered sources) → recommendations + annotated blueprint → action plan (evidence-strength × effort matrix + phased roadmap + the metric) → sources & method. Templates in references/report-structure.md.

Build & cite

• Build with the render-safe component library in references/html-components.md. Obey the rules: never put text inside data-width bars (use label|track|value rows); never use absolute-positioned label clouds (use quadrant grids); collapse every grid on mobile; keep prose out of visuals; match the subject's real brand tokens.
• Generate outline-first, section-by-section (pass prior headers in to avoid duplication), then a polish pass. Cite with the gapless sequential spec and a one-per-line Sources list (references/citations-and-grounding.md).

QA before delivering

Run references/qa-checklist.md (tag balance, node --check on any JS, a render check, scripts/check_citations.py for links, figures and footnotes) plus the grounding gate from references/citations-and-grounding.md (every load-bearing claim cited; numbers/dates attributed; no sources dropped in a merge).

Reference map

• references/report-structure.md — section templates, IA principles, audience calibration, the annotated-blueprint pattern.
• references/html-components.md — render-safe HTML/CSS component library + the rendering-bug rules.
• references/citations-and-grounding.md — citation spec, outline-first generation, question-type templates, tone presets, and the grounding/self-verification gate.
• references/qa-checklist.md — pre-delivery verification steps and commands.
• references/experiment-readout.md — the one-page A/B test readout (numbers from experiment-design/scripts/readout.py).
• scripts/check_citations.py — checks every cited URL resolves and HTML footnotes are gapless; checks load-bearing numbers against their source page when given a --claims CSV (url,expect).

Engagement context

At the start of a task, check for the engagement brief (default ./cro-engagement.md, created by the getting-started skill / /kickoff). If it exists, read it first and match the deliverable's audience (internal vs client-facing) and the brand voice/tokens it records. If it's missing and the work is more than a one-off, offer to capture one via getting-started.

Related skills

deep-research-synthesis (supplies tiered evidence), ecommerce-cro-audit (a frequent input), experiment-design (test readouts), voc-research (customer themes and quotes). Agents: report-builder runs this skill; cro-verifier checks citations and grounding before ship.

Guardrails

Don't put explanatory prose inside visuals. Don't ship a visual that could overlap/clip at any width. Always verify before presenting; offer a browser screenshot if you can't render locally.