AgentSkillsCN

transcript-analysis

解析演讲稿,提取结构化数据

SKILL.md
--- frontmatter
name: transcript-analysis
description: Analyzes transcripts to extract structured data
version: 1.0.0

Transcript Analysis Skill

Purpose

Extract structured information from corrected video transcripts to enable case study generation:

  • Identify CNCF projects used
  • Extract quantitative metrics
  • Classify content into sections

Analysis Tasks

1. Identify CNCF Projects

Find all mentions of CNCF projects and understand their usage context.

Common CNCF Projects:

  • Kubernetes, Prometheus, Envoy, CoreDNS, containerd
  • Fluentd, Jaeger, Vitess, Helm, Argo CD, Flux
  • Cilium, Linkerd, Istio, etcd, CRI-O, Harbor
  • Falco, Dragonfly, Rook, TiKV, gRPC, CNI
  • Knative, OpenTelemetry

For each project found, extract:

  • Project name (exact capitalization)
  • Usage context (what it's used for)
  • Any specific features or benefits mentioned

Example:

json
{
  "name": "Kubernetes",
  "usage_context": "container orchestration and workload scheduling"
}

2. Extract Quantitative Metrics

Find all measurable achievements and improvements.

Metric Types:

Percentages:

  • "50% reduction in..."
  • "3x increase in..."
  • "99.9% uptime"

Time Savings:

  • "from 2 hours to 15 minutes"
  • "deployment time reduced by 30 minutes"
  • "faster by 5x"

Scale:

  • "10,000 pods in production"
  • "1 million requests per second"
  • "100 microservices"

Cost:

  • "$100,000 saved annually"
  • "reduced costs by 40%"
  • "infrastructure costs decreased"

Reliability:

  • "zero downtime deployments"
  • "99.99% availability"
  • "reduced incidents by 80%"

Format for each metric:

json
{
  "value": "50%",
  "type": "percentage",
  "context": "reduction in deployment time",
  "full_statement": "We saw a 50% reduction in deployment time after adopting Argo CD"
}

3. Classify Content into Sections

Analyze the transcript and extract content for each section type.

Section Types:

Background:

  • Company overview and industry
  • Business context and scale
  • Why they're using CNCF technologies
  • Team size and structure

Keywords: "we are", "our company", "we work with", "our team", "in our industry"

Challenge:

  • Problems they faced
  • Pain points and limitations
  • Technical debt or constraints
  • Business pressures

Keywords: "the problem", "we faced", "difficulty", "challenge", "struggled", "couldn't"

Solution:

  • CNCF technologies adopted
  • Implementation approach
  • Architecture changes
  • How they solved problems

Keywords: "we implemented", "we adopted", "we deployed", "we chose", "solution", "approach"

Impact:

  • Results achieved
  • Metrics and improvements
  • Business outcomes
  • Lessons learned

Keywords: "we achieved", "we saw", "improvement", "results", "now we can", "benefit"

Output Format

Return a JSON object with this structure:

json
{
  "cncf_projects": [
    {
      "name": "Kubernetes",
      "usage_context": "container orchestration platform for microservices"
    },
    {
      "name": "Argo CD",
      "usage_context": "GitOps continuous delivery for Kubernetes"
    }
  ],
  "key_metrics": [
    {
      "value": "50%",
      "type": "percentage",
      "context": "reduction in deployment time",
      "full_statement": "We reduced deployment time by 50%"
    },
    {
      "value": "10,000",
      "type": "scale",
      "context": "pods managed in production",
      "full_statement": "We now manage over 10,000 pods in production"
    }
  ],
  "sections": {
    "background": "Relevant sentences and context...",
    "challenge": "Description of problems faced...",
    "solution": "How they implemented CNCF technologies...",
    "impact": "Results and improvements achieved..."
  }
}

Processing Guidelines

  1. Read entire transcript - Understand full context
  2. Identify all CNCF projects - Case-insensitive search
  3. Extract metrics aggressively - Don't miss quantitative data
  4. Classify by strongest signal - Sentences can belong to multiple sections
  5. Preserve original wording - Use actual quotes when possible
  6. Be comprehensive - Include all relevant information

Quality Checklist

  • All CNCF projects identified (minimum 2)
  • Usage context provided for each project
  • At least 1 quantitative metric extracted
  • All 4 section types have content
  • Background explains company context
  • Challenge describes specific problems
  • Solution details CNCF implementation
  • Impact includes measurable results

Example Input

code
We're a financial services company with 5000 employees. We were struggling
with slow deployments that took 2-3 hours. We adopted Kubernetes for
orchestration and Argo CD for continuous delivery. Now our deployments
take only 15 minutes and we manage 10,000 pods across multiple clusters.

Example Output

json
{
  "cncf_projects": [
    {
      "name": "Kubernetes",
      "usage_context": "container orchestration"
    },
    {
      "name": "Argo CD",
      "usage_context": "continuous delivery"
    }
  ],
  "key_metrics": [
    {
      "value": "2-3 hours to 15 minutes",
      "type": "time_savings",
      "context": "deployment time",
      "full_statement": "deployments took 2-3 hours, now take only 15 minutes"
    },
    {
      "value": "10,000",
      "type": "scale",
      "context": "pods managed across clusters",
      "full_statement": "we manage 10,000 pods across multiple clusters"
    }
  ],
  "sections": {
    "background": "We're a financial services company with 5000 employees.",
    "challenge": "We were struggling with slow deployments that took 2-3 hours.",
    "solution": "We adopted Kubernetes for orchestration and Argo CD for continuous delivery.",
    "impact": "Now our deployments take only 15 minutes and we manage 10,000 pods across multiple clusters."
  }
}

Important Notes

  • This analysis feeds into the case-study-generation skill
  • Quality here directly impacts final case study quality
  • Be thorough - missing metrics or projects degrades output
  • When unsure, include rather than exclude
  • Preserve technical accuracy - don't interpret or guess