Skip to content

TsukiNi22/ai-utils / skills/docs/report

report

Verified

Deliver a result (report, audit, review, study, benchmark, summary) as Markdown and/or PDF in the user's report style (measured on the Convertly docs - WeasyPrint, A4, Noto Sans 9 pt, navy (1a3d7c) title with a navy rule under it, navy table heads, light blue callout with navy bar, zebra tables, mono code on grey, footer "title - page / pages"), English by default, converted with scripts/md2pdf.py. Use whenever the user asks for a PDF, a report/benchmark/audit/study, or another skill needs a report output.

Reports in the user's style (Markdown + PDF)

SKILL_DIR = directory of this file. Requires weasyprint + python3-markdown (installed here) and the Noto fonts. The style is templates/report.css: never restyle unless the user asks. Start from templates/report.md.

1. Ask once (AskUserQuestion, French question, per request)

Format du rendu : Markdown + PDF (Recommandé) · PDF seulement · Markdown seulement. Skip the question when the user already said it. Several reports in the same request share the answer.

2. Language and location

  • English by default (titles, tables, text), another language only when the user asks for it.
  • Location: the one given by the user, otherwise the default of the calling skill (audits: audit/ at the root of the repository, or of the current folder outside a repository); never docs/ unless asked.
  • File name: <YYYY-MM-DD>-<subject>.md / .pdf (ex: audit/2026-10-03-quality.pdf).

3. Convert

python3 SKILL_DIR/scripts/md2pdf.py <report.md> --format both|pdf|md [-o report.pdf] [--footer "Project — subject"] [--lang en]

both keeps the .md next to the .pdf; pdf removes the .md after a successful conversion; md writes no PDF.

4. Structure (from the reference documents)

  1. # Project: subject (one H1, underlined by the navy rule; it is also the footer text unless --footer).
  2. Meta line: **Date**: October 3, 2026 · **Sources**: ... (code paths in backticks).
  3. Callout > **How to read this document** + numbered list of the parts + the assumptions / limits / what is not included.
  4. ## Contents: plain numbered list of the H2.
  5. ## 1. Summary: table | Point | Conclusion |, first column in bold, key numbers in bold.
  6. Numbered parts ## N. Title, sub-parts ### N.M Title, tables for every comparison, numbered findings 1. **Finding.** explanation, short > **Key takeaway**: ... callouts.
  7. ## N. Sources: grouped by **Category**: with bullet links <https://...>, then what is an assumption.

5. Writing

  • Factual, short paragraphs, every number sourced or marked as an assumption, units and dates explicit.
  • Tables rather than prose for anything with several attributes; — for empty cells, n/a when not measurable.
  • Code, paths, commands, identifiers in backticks.
  • Variants: {: .warn} / {: .ok} on the line after a callout (orange / green bar), {: .page} after an H2 to start it on a new page. GitHub callouts (> [!NOTE], [!TIP], [!IMPORTANT], [!WARNING], [!CAUTION]) are rendered too (titled, warning / caution orange, tip green): a README or a protocol-doc Markdown converts as is.

6. Check

  • md2pdf.py fits every table by itself (a width per column from its words, smaller, or a landscape page when it is too wide) and tells on stderr a word past the edge of a page: never a width set by hand, never ignored.
  • Convert a page to an image (pdftoppm -r 60 -png -f 1 -l 2 report.pdf /tmp/p) and look at it: title rule present, no table overflowing, lists rendered as lists, footer present.
  • Give the user the path(s) and the page count (pdfinfo).
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 193kUpdated Claude CodeCodexVerifiedInstall