Appendix D: Prompt Templates for Common Tasks#
Data Analysis#
As a Claude prompt (template):
Read the data file at [FILE PATH].
Analyze it and give me:
1. [Specific metric or breakdown]
2. [Another metric]
3. [Comparison or trend]
Create a chart showing [specific visualization].
Save the chart as PNG at 300 DPI to [OUTPUT PATH].
LinkedIn Caption Writing#
As a Claude prompt (template):
[Provide context: file path to source material, or paste key points]
Write a LinkedIn post caption about [topic]. Follow these rules:
- Start with a compelling hook (give me 3 options)
- [Specific formatting rules]
- Tone: [conversational/authoritative/provocative/reflective]
- Length: 800-1200 characters
- Save as RTF to [OUTPUT PATH]
Website Design#
As a Claude prompt (template):
Build a [type of website] with these specifications:
Purpose: [What the site is for]
Pages: [List pages and their content]
Design direction: [Aesthetic, colors, references]
Use the frontend-design skill.
Deploy-ready (static HTML/CSS/JS).
Research Task#
As a Claude prompt (template):
Conduct research on [topic/company/market].
I need to understand:
1. [Research question 1]
2. [Research question 2]
3. [Research question 3]
Use web search and web fetch to gather information from
multiple sources. For each finding, note the source URL.
Save the research as organized files in [FOLDER PATH]:
- raw-findings.md (all gathered information)
- synthesis.md (your analysis and recommendations)
Document Processing#
As a Claude prompt (template):
Read the document(s) at [FILE PATH(S)].
[Specific task: summarize / extract data / compare / convert]
Rules:
- [Formatting rules]
- [What to include/exclude]
- [Output format]
Save to [OUTPUT PATH].
Media Processing#
As a Claude prompt (template):
[Download / Convert / Extract / Transcribe / Merge] the
[video / audio / image] at [FILE PATH or URL].
Specifications:
- [Quality/format requirements]
- [Any processing steps]
- [Output format and location]
Save to [OUTPUT PATH].
Video Download and Processing (Complete Chain)#
As a Claude prompt (template):
Download this YouTube video in the highest available quality:
[YOUTUBE URL]
Then:
1. Extract the audio as a separate MP3 file
2. Transcribe the audio using Whisper
3. Generate an SRT subtitle file from the transcription
4. Burn the subtitles onto the video as a new MP4
5. Create a 5-second GIF from the timestamp [START]-[END]
Save everything to [OUTPUT FOLDER]:
- original-video.mp4
- audio.mp3
- transcript.txt
- subtitles.srt
- video-with-subs.mp4
- clip.gif
Competitive Research Brief#
As a Claude prompt (template):
I need a competitive intelligence brief on [COMPANY/BRAND].
Research the following:
1. Their website: scrape key pages and understand their
positioning, messaging, and product/service offering
2. Their content strategy: what topics do they publish about?
What's their content frequency?
3. Any recent news or announcements
4. How they compare to [OUR COMPANY/APPROACH] on these
dimensions: [list 3-5 dimensions]
Format the output as a brief with:
- Executive summary (3 bullets)
- Positioning analysis
- Content strategy overview
- Strengths and potential gaps
- Opportunities for differentiation
Save as both markdown and RTF to [OUTPUT FOLDER].
Multi-Step Content Pipeline#
As a Claude prompt (template):
I'm creating a data-driven LinkedIn post. Here's the pipeline:
1. Read the dataset at [FILE PATH]
2. Analyze: [specific analysis — e.g., "Which categories have
the highest growth rate year-over-year?"]
3. Create a clean, professional chart showing the top finding
- Use [color scheme] as the palette
- Make it LinkedIn-friendly (square or slightly horizontal)
- Save as PNG at 300 DPI
4. Write a LinkedIn caption:
- Lead with the most surprising data point as the hook
- 3 hook options, each using a different angle
- Reference the specific numbers from the analysis
- Tone: [conversational/authoritative/provocative]
- Length: 800-1200 characters
- Save as RTF
Save all outputs to [OUTPUT FOLDER].
Project Setup (First Session)#
As a Claude prompt (template):
We are starting a new project: [PROJECT DESCRIPTION].
Set up the following:
1. Create a CLAUDE.md with:
- Project overview and goals
- Key deliverables with checklist
- Any rules or constraints
- File structure documentation
2. Create /data/ folder for raw inputs
3. Create /output/ folder for deliverables
4. Create /memory/ folder for supplementary context
5. Create a README.md documenting the project structure
After setup, read the files I've placed in /data/ and give me
an overview of what we're working with.
PDF Report Creation#
As a Claude prompt (template):
Create a professional PDF report using the /document-skills:pdf skill.
Title: [REPORT TITLE]
Purpose: [What this report is for and who will read it]
Include these sections:
1. [Section 1 title] — [brief description of content]
2. [Section 2 title] — [brief description]
3. [Section 3 title] — [brief description]
[etc.]
Source material:
- [FILE PATH to data/research/analysis]
- [Additional context or file paths]
Design:
- Clean, professional layout
- [Color scheme]
- Include page numbers and headers
- Tables should be well-formatted and scannable
Save the PDF to [OUTPUT PATH].
Session Handoff (End of Session)#
As a Claude prompt (template):
Before we end this session, please:
1. Summarize what we accomplished today
2. List any decisions we made and their rationale
3. Note anything that's still pending or in progress
4. Update the CLAUDE.md with:
- New decisions made today
- Updated progress tracker
- Any new rules or learnings
5. If there are open tasks, save a session handoff document
to /memory/ with enough context to resume tomorrow
Confirm all updates are saved.