Iopho Product Context

iopho-product-context

Generate structured context.md for video production — the foundation file all other iopho skills read.

When to Use

• Starting a new video project (product demo, launch video, explainer, VSL)
• Onboarding a client project (like a client project)
• Any iopho skill asks for context but none exists yet
• Updating an existing context file with new information

Quick Start

/iopho-product-context ./exp/my-project
/iopho-product-context ./exp/my-project --quick           # 5 core questions only
/iopho-product-context ./exp/my-project --from-file README.md  # pre-fill from docs

How It Works

Mode 1: Interactive (default)

Ask the user these questions IN ORDER. Accept answers in any language (CN/EN/mixed).

Core Questions (always ask):

1. What's the product? — Name, URL, one-sentence description
2. Who is it for? — Primary audience + their pain point
3. What makes it different? — Top 3 differentiators from competitors
4. What video do you want? — Type (demo/explainer/launch/teaser), duration, platforms
5. Any brand rules? — Colors, fonts, tone of voice, design system files

Deep-Dive Questions (ask unless --quick):

1. Who are the competitors? — List 2-5 competitors, their strengths/weaknesses
2. Founder visibility? — Should founders/team appear? (public / anonymous / team-only)
3. Hero feature? — The ONE feature that must be the centerpiece of the video
4. What to avoid? — Anything off-limits (features, messaging, visuals)
5. Reference videos? — Any video styles they like (provide URLs → feed to iopho-searching-videos)
6. Languages? — Which languages for VO and subtitles
7. Call to action? — What should viewers do after watching

Mode 2: Auto-Detect (--from-file)

Before asking questions, scan the project directory for existing context:

Look for and read (if present):
├── README.md              → product name, description, features
├── package.json           → name, description, keywords
├── **/brand.ts            → colors, fonts (Remotion projects)
├── **/*DesignPrompts*     → design system (client project pattern)
├── **/questionaire.md     → existing intake (client project pattern)
├── **/context.md          → previous context file (update mode)
└── website URL            → scrape tagline, features, pricing (if provided)

Pre-fill answers from detected files, then ask user to CONFIRM or CORRECT each field. Skip questions already answered by auto-detection.

Mode 3: Update

If context.md already exists in the project directory, read it and ask: > "Found existing context.md. What needs updating?"

Only modify the fields the user mentions. Preserve everything else.

Output

Generate {project-dir}/context.md using the template at templates/context-template.md.

Output rules:

• YAML frontmatter: machine-readable fields (name, URL, colors as hex, platforms as array)
• Markdown body: human-readable prose (features, audience, competitors, notes)
• All hex colors validated (6-char format)
• All URLs as full https:// format
• Bilingual content preserved as-is (don't translate user's words)

Validation

After generating context.md, verify:

• [ ] Product name is not empty
• [ ] At least 1 target audience defined
• [ ] At least 1 video platform specified
• [ ] At least 1 language specified
• [ ] Brand colors are valid hex (if provided)
• [ ] Founder visibility preference is set

If any critical field is missing, ask the user — don't guess.

Real-World Examples

[App Name] (a reader/learning app):

• Product: AI read-later app → Explainer video, 78s, YouTube/Product Hunt
• Brand: #007AFF primary, SF Pro font, warm/approachable tone
• Hero feature: AI reflow + instant summary
• Context existed in: {project}/video-strategy.md

[CLI Tool] (a developer networking tool):

• Product: P2P agent collaboration → Launch video, 90s, YouTube/GitHub
• Brand: #1D3557 deep ocean, #E63946 lobster red, cinematic/authoritative
• Hero feature: ASCII globe + npx install
• Context existed in: {project}/questionaire.md + {project}/product-codex.md
• Founder visibility: anonymous (user requested)

Related Skills

• /iopho-video-director — calls this skill as Phase 0 step 1
• /iopho-searching-videos — uses reference video URLs from context
• /vsl-storyboard-writer — reads context for audience/positioning
• /storyboard — reads context for product understanding

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: iopho-team
• Source: iopho-team/iopho-skills
• License: MIT

Install and usage instructions live in the source repository linked above.