presentation-evidence
Skill0 starsChanged 24 days agoMITView on GitHub
Builds a traceable claim ledger from the brief and source inventory, then classifies what is safe to say in the main talk, notes, backup, or not at all. Use when the evidence gate is current for a presentation project or when a router asks for a supported claim ledger.
In this repository42 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .agents/skills/vercel-composition-patterns/AGENTS.mdAGENTS.md
- .agents/skills/vercel-react-best-practices/AGENTS.mdAGENTS.md
- apple-designSkill
- apply-commentsSkill
- create-slideSkill
- create-themeSkill
- current-slideSkill
- diagram-designSkill
- emil-design-engSkill
- frontend-designSkill
- genofficeSkill
- grillingSkill
- grill-meSkill
- human-writingSkill
- make-presentationSkill
- presentation-deck-qaSkill
- presentation-evidenceSkill
- presentation-product-grillSkill
- presentation-prototypeSkill
- presentation-source-intakeSkill
- presentation-stage-grillSkill
- presentation-story-architectureSkill
- presentation-story-grillSkill
- review-animationsSkill
- shadcnSkill
- slide-authoringSkill
- to-questionnaireSkill
- vercel-composition-patternsSkill
- vercel-react-best-practicesSkill
- wait-whatSkill
- wang-yongfu-teaching-craftSkill
- web-design-guidelinesSkill
- genofficeSkill
- make-presentationSkill
- presentation-deck-qaSkill
- presentation-product-grillSkill
- presentation-prototypeSkill
- presentation-source-intakeSkill
- presentation-stage-grillSkill
- presentation-story-architectureSkill
- presentation-story-grillSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 5 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Use when evidence is the current gate.
- Very specific: line 10
Read repo instructions, PRESENTATION-BRIEF.md, SOURCE-INVENTORY.md, and the cited source files before writing new claims.
- Very specific: line 11
Read the current mode from STATUS.md.
- Very specific: line 14
Write EVIDENCE-LEDGER.md from templates/EVIDENCE-LEDGER.md.
- Distinctive: line 15
Create one row per material claim, not one row per source file.
- Distinctive: line 17
Add dates, definitions, units, or methods when they are necessary to keep a material claim honest.
- Very specific: line 18
Use references/claim-safety.md for research-enhanced claim safety.
- Distinctive: line 23
Explicit Baselines & Denominators: Always capture the exact reference year, definition, denominator, and global baseline comparisons (e.g.
- Distinctive: line 36
Add backup/reference safeguards for contested, overflow, or Q&A-only material.
- Distinctive: line 40
Stop once every material claim is supported, carefully qualified, moved to notes or backup, or rejected.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Build the evidence ledger
Use when evidence is the current gate.
- Read repo instructions,
PRESENTATION-BRIEF.md,SOURCE-INVENTORY.md, and the cited source files before writing new claims. - Read the current mode from
STATUS.md; do not assume the mode selected at intake is still current. - General mode keeps evidence proportionate.
- Record provenance and sufficient qualification for material claims.
- Write
EVIDENCE-LEDGER.mdfromtemplates/EVIDENCE-LEDGER.md. - Create one row per material claim, not one row per source file.
- Record the exact source, source type, caveat, allowed use, and the slide or range the claim may support.
- Add dates, definitions, units, or methods when they are necessary to keep a material claim honest.
- Use
references/claim-safety.mdfor research-enhanced claim safety.
Authority & scientific fidelity rules
- Primary & Official Authority Preference: For demographic, economic, suicide/mortality, birth rate, or public health claims, mandate primary/official sources (e.g. official statistical bureaus like DSEC, WHO, World Bank, peer-reviewed journals) rather than secondary news reports or casual commentary.
- Explicit Baselines & Denominators: Always capture the exact reference year, definition, denominator, and global baseline comparisons (e.g. regional suicide rate vs WHO global average).
- Nuanced Literature Fidelity: Guard against oversimplified pop-clichés (e.g. do not reduce complex findings like the Killingsworth/Kahneman income-wellbeing association to "money can't buy happiness").
Research-enhanced mode alone adds A/B/C/D claim safety.
A: directly supportable in the main talk.B: reasonable interpretation with careful wording.C: contested or context-dependent, notes or backup only.D: unsafe or unsupported, reject.- Record exact dates, definitions, units, denominators, and methods.
- Record correlation-versus-causation review.
- Record competing measures and stale-value checks.
- Apply official and primary-source preference before secondary summaries.
- Add backup/reference safeguards for contested, overflow, or Q&A-only material.
- Move unsupported but useful material to notes or backup only when the wording stays honest.
- Reject claims that cannot be supported proportionately for the presentation’s mode.
- Stop once every material claim is supported, carefully qualified, moved to notes or backup, or rejected.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

