presentation-engine — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited presentation-engine (Agent Skill) and scored it 96/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 1 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Orchestration Log: When this skill is activated, append a log entry tooutputs/orchestration_log.md: ``### Skill Activation: Presentation Engine **Timestamp:** [current date/time] **Actor:** AI Agent (presentation-engine) **Input:** [paper source], target format: [conference talk / workshop / poster] **Output:** presentation.md with [N] slides + speaker notes``
A conference presentation is NOT a paper read aloud. It's a persuasion exercise: convince the audience in 15-20 minutes that your research matters, your method is sound, and your findings are interesting. This engine transforms a paper into a focused, visually driven presentation with clear narrative flow.
/generate-slidesdraft.md or latex/paper.tex exists (complete paper)figures/ directory with generated figures (optional but recommended)Read the paper and extract:
| Section | Extract | Max Slides |
|---|---|---|
| Title/Abstract | Title, authors, affiliations, 1-sentence summary | 1 |
| Introduction | Motivation hook, gap, RQ(s) | 2-3 |
| Theoretical Background | Key theory/framework, 1-2 core concepts | 1-2 |
| Methodology | Method name, key design choices, data summary | 2-3 |
| Results/Findings | Top 3-5 findings, key tables/figures | 4-6 |
| Discussion | Key implications (2-3), limitations (1-2) | 2-3 |
| Conclusion | Main takeaway, future directions | 1 |
## Slide Structure
1. **Title Slide** — Title, authors, affiliations, venue, date
2. **Motivation** — Why should the audience care? (practical hook)
3. **Problem / Gap** — What don't we know? What's missing?
4. **Research Question(s)** — Clear, prominent, memorable
5. **Theoretical Lens** — Key framework (1 visual if possible)
6. **Method Overview** — Approach in 3-5 bullet points + visual
7. **[Method Detail]** — Optional: sample, process, timeline
8. **Results 1** — First major finding (with figure/table)
9. **Results 2** — Second major finding
10. **Results 3** — Third major finding (or synthesis)
11. **[Results 4]** — Optional: additional finding or robustness
12. **Discussion: What This Means** — Interpretation of findings
13. **Implications** — For theory AND practice (split or combined)
14. **Limitations & Future** — Honest, brief, forward-looking
15. **Conclusion / Takeaway** — One memorable message
16. **Thank You / Q&A** — Contact info, paper referenceRemove slides 6 (method detail), 11 (results 4), and merge 12-13.
Add: deeper literature context, method details, additional findings, more discussion, interactive elements.
Generate slides in Marp-compatible markdown format:
---
marp: true
theme: academic
paginate: true
---
# [Paper Title]
## [Subtitle if applicable]
**[Author 1]**, [Author 2], [Author 3]
[Affiliation(s)]
[Venue] — [Date]
---
# Motivation
[Striking opening statement or statistic]
- [Key trend 1]
- [Key trend 2]
- [Key challenge]
> "[Compelling quote or statistic]" — [Source]
<!-- speaker notes:
Open with the practical relevance. Spend 1 minute here.
Make the audience feel the problem before you state the RQ.
-->
---
# The Gap
What we **know**: [established knowledge, 2-3 points]
What we **don't know**: [specific gap]
Why it **matters**: [consequence of the gap]
<!-- speaker notes:
Transition from what we know to what's missing.
Emphasize why this gap matters for both theory and practice.
-->
---
# Research Question
> **RQ: [Research Question]**
[1-2 sentences contextualizing the question]
<!-- speaker notes:
Pause after showing the RQ. Let the audience read it.
This is the anchor of your entire talk.
-->
---For each slide:
For figures from the paper:
figures/fig_*.png filesFor Motivation slides:
Open with [the hook]. Pause for effect.
Key message: [what the audience should feel/understand].
Transition: "This brings us to the question of..."
Time: ~1 minuteFor Results slides:
Present [finding]. Point to [specific part of figure/table].
Explain what this means: [interpretation in plain language].
Connect to RQ: "This tells us that [answer to RQ aspect]."
If asked about [likely question]: [prepared answer].
Time: ~2 minutes per result slideFor Discussion slides:
Summarize: "So what does this mean?"
Key implication: [most important takeaway].
Compare to prior work: "Unlike [Author], we find [contrast/extension]."
Time: ~2 minutes| File | Content |
|---|---|
presentation/presentation.md | Marp-compatible slide deck with speaker notes |
presentation/speaker_notes.md | Standalone speaker notes (for rehearsal) |
presentation/slide_outline.md | Slide-by-slide outline (for quick reference) |
presentation/figures/ | Presentation-optimized figures (if modified) |
🎤 PRESENTATION READY
━━━━━━━━━━━━━━━━━━━━
Paper: "[Title]"
Format: [Conference talk / Workshop / Seminar]
Slides: [N] slides (~[N] minutes estimated)
Structure:
1-3: Introduction & Motivation
4-5: Theory & Method
6-10: Results
11-14: Discussion & Implications
15-16: Conclusion & Q&A
📁 Saved: presentation/
→ Review slides and speaker notes
→ Rehearse with speaker notes (target: [N] min)
→ Render with Marp: marp presentation.md --pdf| Scenario | Integration |
|---|---|
| After paper acceptance | Natural next step |
| Figure-engine | Generate presentation-specific visuals |
| Writing-engine | Refine speaker note language |
| Submission-engine | Presentation deadline tracking |
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.