CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

jbaruch/speaker-toolkit

Twelve skills for conference speakers covering rhetoric analysis, talk authoring, deck building and export, CFP submissions, publishing, illustrations, shownotes, screencast production, and vertical shorts, informed by a 113-entry Presentation Patterns taxonomy (83 observable: 64 patterns + 19 antipatterns; 30 unobservable: 21 patterns + 9 antipatterns).

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Medium

Suggest reviewing before use

Overview
Quality
Evals
Security
Files

editorial-review.mdskills/presentation-creator/references/

Draft and Review Toolkit Text

Start with the speaker and the material

Read approved examples in the same medium when available. Use delivered remarks for notes, approved decks for headlines, and approved submissions for CFPs. Pick a few passages that show how the speaker makes a claim, tells a story, or lands a joke. Match each speaker to their own examples in a multi-speaker talk. Use the vault’s confirmed intent to interpret the examples. When examples are missing, write in a plain register and leave voice calibration unresolved.

Identify the reader and what the artifact needs to convey. Write from the talk’s claim, examples, and sequence. Preserve author-supplied wording, approved copy, quoted material, technical meaning, and the format each consumer expects.

Edit for the medium

  • Narrative and outline. State what happens and why it changes the argument. Make each chapter beat follow from the preceding one. Replace planning shorthand and generic transitions with the claim, event, or example that earns the turn.
  • Speaker notes, rehearsal lines, and demos. Read the lines aloud in sequence. Use phrases the assigned speaker would say. Give a demo line something concrete to point at. Keep stage directions and build cues in their own fields.
  • Slides and short hooks. Read headlines in sequence and against their visuals. Put a claim, instruction, or punchline on screen when it has a job there. A pun should make the point easier to remember. Use the slide-copy examples in phase3-content.md.
  • Build instructions, analyses, and reviews. Name the object, action, and relevant evidence. Say which slide, line, or behavior needs attention. Use plain production language. Keep source quotes, measured data, and validation results intact.
  • CFPs, shownotes, and other public prose. Ground the promise in what the talk actually delivers. A proposal tells the committee what the audience will learn. Shownotes acquaint readers with the actual argument and examples. A downloadable skill explains enough of the talk to answer questions about it.

Before presenting a draft or asking for approval, cut generic openings, repeated claims, and transitions that add no information. Check the text beside the source and revise the draft where it loses the speaker’s point or specificity. Fix authored fields in the owning artifact. Regenerate and read the derived outputs as a reader would, including the partial narrative and, when present, the full narrative, rehearsal script, build sheet, and rhetorical review. If the boilerplate comes from a renderer, flag it in the draft review and leave the generated file read-only. Preserve fixed labels and machine-readable syntax.

For a demo about skill discovery, the narrative might say “Give the agent the same request twice. Change only the skill’s description between runs.” The speaker note can point at the cause. “See that line? That’s the only thing I changed.” The build sheet can tell the operator what to show. “Put the descriptions side by side. Highlight the trigger phrase after the second run.”

Use the scoped blog-writer review

For a new or substantively revised draft, batch its changed prose in reading order with enough neighboring text to judge transitions. When the installed skill supports scoped non-blog review, invoke Skill(skill: "blog-writer") for a scoped, review-only humanization pass on non-blog text. Supply the reader, purpose, field or slide locations, approved wording, and relevant voice examples. Ask for the completed manual catalog review and sweep, with each finding’s location, phrase, suggested edit, and disposition. For the presentation guardrail, also ask for severity (high or medium) on confirmed findings so they can be counted. Keep the scope to the supplied prose. Leave identifiers, syntax, raw quotes, transcripts, numerical results, and visual prompts out of the scan.

When drafting or revising, review suggestions in context and apply warranted edits before approval. Keep intentional phrasing when a source passage or explicit direction supports it in this context. Check revised passages for meaning, voice, and continuity. Run scoped review again on passages you substantially rewrite. Leave approval and persistence with the calling skill.

For a guardrail scan or other review-only task, report findings and suggestions. Leave the draft text untouched.

If the scoped review is unavailable, fails, or requests a full blog workflow, continue the toolkit’s editorial pass and report the reason the scan was skipped. For a missing or older install, point to tessl install jbaruch/blog-writer. Keep the anti-pattern catalog and sweep in their owning skill.

skills

presentation-creator

references

adapting.md

editorial-review.md

idea-sourcing-wizard.md

outline-artifacts.md

outline-handoff.md

pattern-history-authorization.md

pattern-menu.md

phase0-intake.md

phase1-intent.md

phase2-architecture.md

phase3-content.md

phase4-guardrails.md

runtime-and-context.md

SKILL.md

README.md

tile.json