CtrlK
BlogDocsLog inGet started
Tessl Logo

jbaruch/speaker-toolkit

Four-skill presentation system: ingest talks into a rhetoric vault, run interactive clarification, generate a speaker profile, then create new presentations that match your documented patterns. Includes an 88-entry Presentation Patterns taxonomy for scoring, brainstorming, and go-live preparation.

96

1.21x
Quality

93%

Does it follow best practices?

Impact

97%

1.21x

Average score across 30 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Overview
Quality
Evals
Security
Files

criteria.jsonevals/scenario-11/

{
  "context": "Tests whether the agent correctly applies the skill's error handling matrix (transcript/slides success/fail to correct outcome categories), slide source determination logic, skip-for-missing-sources rules, and additive summary update policy.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Full success categorized correctly",
      "description": "Talks where both transcript and slides succeeded (Microservices Myths, Platform Engineering) get a status indicating full successful processing — distinguishable from partial or failed outcomes",
      "max_score": 10
    },
    {
      "name": "Transcript fail + slides OK = partial",
      "description": "Talks where transcript failed but slides succeeded (Security Shift Left) get a status indicating partial processing — distinguishable from both full success and full failure",
      "max_score": 10
    },
    {
      "name": "Slides fail + transcript OK = partial",
      "description": "Talks where transcript succeeded but slides failed (DevOps Culture Reset) get the same partial status category as the transcript-fail case — both are partial outcomes",
      "max_score": 10
    },
    {
      "name": "Both fail = skip/fail status",
      "description": "Talks where both transcript and slides failed (GitOps Journey) get a status indicating download failure — distinguishable from partial processing and from the missing-sources skip",
      "max_score": 10
    },
    {
      "name": "Missing sources = distinct skip",
      "description": "Container Security Deep Dive (no video_url, no slides_url, no pptx_path) gets a skip status that is distinct from the download-failure status — it was never attempted, not failed",
      "max_score": 10
    },
    {
      "name": "No video = not processable",
      "description": "Testing in Production (has slides_url but no video_url) is treated as not processable — a talk requires video_url to enter the processing pipeline, regardless of other sources",
      "max_score": 8
    },
    {
      "name": "Slide source set to pptx",
      "description": "Platform Engineering (has pptx_path but no slides_url) gets slide_source indicating PPTX as the source",
      "max_score": 8
    },
    {
      "name": "Slide source set to both",
      "description": "Talks with both slides_url and pptx_path (Microservices Myths, GitOps Journey) get slide_source indicating both sources are available",
      "max_score": 8
    },
    {
      "name": "Slide source set to pdf",
      "description": "Talks with only slides_url and no pptx_path (DevOps Culture, Security Shift Left) get slide_source indicating PDF as the source",
      "max_score": 8
    },
    {
      "name": "Summary additive only",
      "description": "The updated rhetoric summary preserves ALL existing content from the original (Section 1, 2, 4, 7, 15 text is still present) — nothing was deleted or replaced",
      "max_score": 10
    },
    {
      "name": "Summary has new content",
      "description": "The updated rhetoric summary adds new observations from the successfully processed talks (at least one new pattern or refinement added)",
      "max_score": 8
    }
  ]
}

evals

README.md

tile.json