agentskills.codes
TD

tda-paper-dissemination-pack

Use only after a TDL paper section or result set is stable — to convert it into public-facing web pages, talks, posters, slide decks, graphical abstracts, or explainer material.

Install

mkdir -p .claude/skills/tda-paper-dissemination-pack && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/14551" && unzip -o skill.zip -d .claude/skills/tda-paper-dissemination-pack && rm skill.zip

Installs to .claude/skills/tda-paper-dissemination-pack

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Use only after a TDL paper section or result set is stable — to convert it into public-facing web pages, talks, posters, slide decks, graphical abstracts, or explainer material.
177 charsno explicit “when” trigger

About this skill

TDA Paper Dissemination Pack

Package stable research outputs for public consumption: zktheory.org pages, talks, posters, graphical abstracts, blog companions. Strictly downstream of review and verification — dissemination never runs ahead of the evidence.

Tier 3 Constraint

This skill may generate ideas, prototypes, plans, or communication artifacts. It may not create paper claims, result artifacts, canonical computations, or contract-bearing implementations unless routed through the relevant tier 1 or tier 2 skill first.

Do Not Use When

  • The result is provisional (PROVISIONAL flags travel with results into every derived artifact — a poster of a provisional result is a provisional poster nobody will read as provisional).
  • Literature sources are unverified.
  • Paper claims are still under active review or revision.
  • Figures lack regeneration scripts.
  • The output would overstate significance — proportionality applies harder in public material, not softer.

Required Inputs

Stable paper draft or section · verified citations · regenerable figures · paper target · intended audience · claims allowed · claims prohibited. The allowed/prohibited claim lists come from the paper's claim tracing, not from memory.

Procedure

  1. Confirm stability: which draft version, which result files, any PROVISIONAL flags outstanding?
  2. Fix the audience and format (web page / talk / poster / abstract / explainer).
  3. Build the claims budget: what this artifact may assert, at what strength, and what it must not touch (e.g. P01-B material never carries applied poverty interpretation).
  4. Reuse paper figures via their regeneration scripts; public simplifications of figures are new figures and get the same provenance treatment.
  5. Simplify language without shifting claim strength — "suggests" does not become "shows" on a poster.
  6. Route the draft artifact through a prose/communication review pass before publication; anything on the public web is effectively permanent.

Self-Test Prompts

  • A talk slide is drafted from a result still flagged PROVISIONAL pending the frozen-loadings reruns. → Expected: refuse or mark the slide provisional-and-internal; public material waits for the rerun.
  • A website page rounds p = 0.003 to "conclusive evidence". → Expected: proportionality violation; restate at the statistic's strength.

Escalate Or Stop When

  • Dissemination timing interacts with journal policy (preprint rules, embargoes, the P04-after-arXiv sequencing) — User decision.

Related Skills

tda-peer-review-panel + tda-literature-verification + tda-visualisation-and-diagramming (all upstream gates) · paper-claim-trace (the claims budget) · humanizer (public prose pass).

Search skills

Search the agent skills registry