Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Convert markdown to DOCX, PPTX, XLSX, PDF office documents — use when you need exportable deliverables
.claude/skills/hashgraph-online-skill-doc-delivery/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 68% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 85% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 75% | 0% |
> Host: Codex CLI — This skill was designed for Claude Code and adapted for Codex. > Cross-reference commands use installed skill names in Codex rather than /octo:* slash commands. > Use the active Codex shell and subagent tools. Do not claim a provider, model, or host subagent is available until the current session exposes it. > For host tool equivalents, see skills/blocks/codex-host-adapter.md.
Convert knowledge work outputs from markdown to professional office documents (DOCX, PPTX, XLSX).
This skill helps you transform knowledge work results into polished, deliverable documents:
Before converting documents, verify the document-skills plugin is installed:
bash/plugin list | grep document-skills
If not installed:
bash/plugin install document-skills@anthropic-agent-skills
Best for PPTX (PowerPoint):
Best for DOCX (Word):
Best for PPTX (PowerPoint):
Best for DOCX (Word):
Best for DOCX (Word):
Best for PDF:
Knowledge work outputs are stored in:
bash~/.claude-octopus/results/
List recent outputs:
bashls -lht ~/.claude-octopus/results/ | head -10
Ask yourself:
The document-skills plugin provides these capabilities:
For DOCX (Word):
Use the /document-skills:docx skill to convert markdown to Word format.
Supports headings, lists, tables, and formatting.For PPTX (PowerPoint):
Use the /document-skills:pptx skill to convert markdown to PowerPoint.
Each ## heading becomes a slide, bullet points auto-format.For PDF:
Use the /document-skills:pdf skill to generate PDF documents.
Ideal for final deliverables and archival.After conversion, consider:
User ran one workflow, wants one document:
User: "Export my latest synthesis to a Word document"
1. Check ~/.claude-octopus/results/ for most recent .md file
2. Identify it's from synthesize workflow
3. Recommend DOCX for academic report format
4. Use document-skills:docx to convert
5. Save to appropriate locationUser wants to create a deck from research:
User: "Create a PowerPoint from this research"
1. Locate the research markdown
2. Identify key sections (## headings)
3. Use document-skills:pptx to convert
4. Each ## heading becomes a slide
5. Recommend adding title slide and summaryUser wants multiple formats:
User: "Create both a Word doc and PowerPoint from this strategy"
1. Locate source markdown
2. Convert to DOCX using document-skills:docx
3. Convert to PPTX using document-skills:pptx
4. Provide both file pathsmarkdownInput: ~/.claude-octopus/results/empathize-session-2026-01-18.md Output Goal: Stakeholder presentation Steps: 1. Read the empathize markdown 2. Extract persona sections 3. Convert to PPTX using document-skills:pptx 4. Each persona becomes a slide 5. Add title slide and key insights summary 6. Save as "UX-Personas-2026-01-18.pptx"
markdownInput: ~/.claude-octopus/results/advise-market-analysis-2026-01-18.md Output Goal: Comprehensive business case document Steps: 1. Read the advise markdown 2. Structure as: Executive Summary, Analysis, Recommendations, Appendix 3. Convert to DOCX using document-skills:docx 4. Add table of contents 5. Format tables and charts properly 6. Save as "Market-Analysis-Business-Case-2026-01-18.docx"
markdownInput: ~/.claude-octopus/results/synthesize-ai-safety-2026-01-18.md Output Goal: Academic paper format Steps: 1. Read the synthesize markdown 2. Structure as: Abstract, Introduction, Literature Review, Gaps, Conclusions 3. Convert to DOCX using document-skills:docx 4. Apply academic formatting (citations, references) 5. Export to PDF for final version 6. Save as "AI-Safety-Literature-Review-2026-01-18.pdf"
If user requests export but no workflow was run:
Response: "I don't see any recent knowledge work outputs. Would you like to run:
- /octo:empathize (UX research)
- /octo:advise (strategic analysis)
- /octo:synthesize (literature review)
Then I can convert the results to your preferred format."If user doesn't specify format:
Response: "I can convert this to:
- DOCX (Word) - Best for detailed documentation
- PPTX (PowerPoint) - Best for presentations
- PDF - Best for final deliverables
Which format would you prefer based on your use case?"If multiple outputs exist:
Response: "I found several knowledge work outputs:
1. empathize-personas-2026-01-15.md (3 days ago)
2. advise-market-analysis-2026-01-18.md (today)
3. synthesize-research-2026-01-10.md (8 days ago)
Which would you like to convert?"If plugin missing:
Response: "The document-skills plugin is required for conversion.
Install it with:
/plugin install document-skills@anthropic-agent-skills
Then I can convert your knowledge work to office formats."This skill is designed to work seamlessly with the three knowledge mode workflows:
Knowledge Work Flow:
1. Run workflow: /octo:empathize (or advise/synthesize)
2. Review markdown output in ~/.claude-octopus/results/
3. Request conversion: "Export to PowerPoint"
4. This skill is used after explicit invocation
5. Professional document deliveredbash# List recent knowledge work outputs ls -lht ~/.claude-octopus/results/ | head -5 # Check document-skills installed /plugin list | grep document-skills # Install document-skills /plugin install document-skills@anthropic-agent-skills # View specific markdown cat ~/.claude-octopus/results/[filename].md
For questions about:
Document delivery skill for claude-octopus v7.3.0+
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | fail→fail | 11,985 | 15,499 | +29% | 1 | 1 | 0% | 2,177 | 3,452 | +59% | 0 | 0 | — |
case-01 | fail→fail | 11,507 | 11,089 | -4% | 1 | 1 | 0% | 1,086 | 2,729 | +151% | 0 | 0 | — |
case-02 | fail→fail | 67,895 | 10,832 | -84% | 1 | 1 | 0% | 7,860 | 2,724 | -65% | 0 | 0 | — |
case-03 | fail→fail | 14,525 | 11,523 | -21% | 1 | 1 | 0% | 2,423 | 2,667 | +10% | 0 | 0 | — |
case-04 | pass→pass | 14,250 | 13,439 | -6% | 1 | 1 | 0% | 2,366 | 3,796 | +60% | 0 | 0 | — |
case-05 | pass→pass | 16,816 | 16,607 | -1% | 1 | 1 | 0% | 2,036 | 4,359 | +114% | 0 | 0 | — |
case-06 | pass→pass | 16,183 | 21,046 | +30% | 1 | 1 | 0% | 2,418 | 4,860 | +101% | 0 | 0 | — |
case-07 | fail→pass | 10,220 | 12,270 | +20% | 1 | 1 | 0% | 1,827 | 3,072 | +68% | 0 | 0 | — |
case-08 | fail→pass | 14,414 | 4,185 | -71% | 1 | 1 | 0% | 1,674 | 3,100 | +85% | 0 | 0 | — |
case-09 | fail→pass | 17,211 | 6,858 | -60% | 1 | 1 | 0% | 2,044 | 2,614 | +28% | 0 | 0 | — |
case-10 | fail→pass | 10,268 | 2,787 | -73% | 1 | 1 | 0% | 1,818 | 2,855 | +57% | 0 | 0 | — |
case-11 | pass→pass | 17,027 | 7,004 | -59% | 1 | 1 | 0% | 1,885 | 3,512 | +86% | 0 | 0 | — |
case-12 | fail→pass | 15,156 | 4,862 | -68% | 1 | 1 | 0% | 1,796 | 3,139 | +75% | 0 | 0 | — |
case-13 | pass→pass | 14,726 | 10,883 | -26% | 1 | 1 | 0% | 1,576 | 3,361 | +113% | 0 | 0 | — |
case-14 | fail→pass | 13,295 | 7,911 | -40% | 1 | 1 | 0% | 2,247 | 3,658 | +63% | 0 | 0 | — |
case-15 | fail→pass | 19,251 | 11,275 | -41% | 1 | 1 | 0% | 2,334 | 3,513 | +51% | 0 | 0 | — |
case-16 | pass→pass | 16,505 | 9,475 | -43% | 1 | 1 | 0% | 1,888 | 3,067 | +62% | 0 | 0 | — |
case-17 | fail→pass | 14,791 | 4,901 | -67% | 1 | 1 | 0% | 1,658 | 3,058 | +84% | 0 | 0 | — |
case-18 | fail→fail | 14,350 | 11,063 | -23% | 1 | 1 | 0% | 1,540 | 2,693 | +75% | 0 | 0 | — |
case-20 | fail→fail | 12,856 | 4,405 | -66% | 1 | 1 | 0% | 1,275 | 2,782 | +118% | 0 | 0 | — |
case-21 | pass→pass | 22,625 | 12,923 | -43% | 1 | 1 | 0% | 2,929 | 3,925 | +34% | 0 | 0 | — |
case-22 | pass→pass | 8,936 | 2,668 | -70% | 1 | 1 | 0% | 1,424 | 2,762 | +94% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +36 percentage points is the difference between those two pass rates over the 18 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.