kit-pm-skills

Query or modify Kit account data via the API. Use this whenever you want to look something up in Kit, pull subscriber data, check stats, create or update records, test an endpoint, or interact with the Kit API in any way — even if you don't know the exact endpoint needed.

Research and produce a competitor analysis for any feature area or topic. Use this when you need competitive context for a PRD, are exploring a new feature area, responding to competitive pressure, or want to understand how competitors approach a specific problem. Even a vague topic works — describe the area and this command will identify relevant competitors, research their current state, and produce an insight-driven analysis following Kit's style guide. Use any time the user mentions "how do competitors handle X", "what does Beehiiv/Mailchimp/ActiveCampaign do with X", "competitive analysis", or "where does Kit stand on X compared to the market".

Draft a Knowledge Base briefing for the support/docs team for a new feature or update

Create a PRD whenever you're scoping a new feature, writing up an idea, or documenting a product direction. Also use to revise an existing PRD by passing a Linear ticket ID (e.g. "refine ECO-123" or "flesh out PROD-76"). Use this any time the user mentions writing a spec, documenting a feature, exploring a product idea, or wanting to turn a rough Linear ticket into a proper requirements doc.

Build an interactive product prototype — researches the feature area, generates an overview page and interactive prototype, and pushes to prototypes.kit.com. Run any time you want to create a new clickable prototype to share with design, engineering, or leadership.

Use this agent when you need to review HTML prototype files for correctness before publishing. Specialised in checking Kit prototype files for broken relative paths, missing files, basic accessibility, and browser console errors. Use after building prototype.html and index.html, before pushing to GitHub Pages. Examples: - Example 1: user: 'Review the prototype files before pushing' assistant: 'I'll have the code-reviewer check both HTML files for broken paths and quality issues.' - Example 2: user: 'Make sure there are no 404s in the new prototype' assistant: 'Let me spin up the code-reviewer to audit the relative paths in both files.'

Use this agent when the user needs to execute Kit API requests, write scripts to interact with the Kit API, handle OAuth token management, parse and format API responses, or debug API call failures. This is the agent for anything involving curl commands, API authentication, or Kit-specific scripting. Examples: - User: "List all my subscribers" Assistant: "I'll use the coder agent to make the API request and format the results." - User: "Bulk tag these subscribers" Assistant: "I'll use the coder agent to run the OAuth-authenticated bulk endpoint." - User: "My OAuth token expired" Assistant: "I'll use the coder agent to refresh the token and retry the request." - User: "Create a subscriber and tag them in one go" Assistant: "I'll use the coder agent to chain the API calls correctly."

Use this agent when the user asks to polish, edit, refine, proofread, or improve written text. This includes PRDs, Slack messages, emails, documentation, announcements, release notes, or any draft that needs professional editing. Also use when the user shares a block of text and asks for feedback on clarity, tone, or grammar. Examples: - User: "Can you clean up this PRD section for me?" Assistant: "Let me use the copywriter agent to refine this." - User: "I need to send this email to the engineering team, can you make it sound better?" Assistant: "I'll use the copywriter agent to polish this email." - User: "Here's a draft announcement about our new API changes. Tidy it up." Assistant: "I'll pass this to the copywriter agent to sharpen the tone and clarity."

Use this agent when you need to research Kit's current state for a feature area — checking internal docs, PRDs, feature notes, and competitor analysis for existing context before making a product decision. This agent searches the workspace and Kit's developer documentation to surface what's already known, so you don't start from scratch or contradict prior decisions. Best for: internal research as part of PRD drafting, prototype planning, competitor analysis, or any time you need to know what Kit currently does or has previously decided. Examples: - Example 1: user: 'What do we know about the App Store and how it currently works?' assistant: 'I'll use the kit-knowledge-curator to search the workspace for PRDs, feature docs, and prior analysis on the App Store.' - Example 2: user: 'Before writing this prototype, find existing context on visual automations' assistant: 'Let me have the kit-knowledge-curator surface any PRD excerpts, feature docs, or research we already have.' - Example 3: user: 'What has already been decided about the sequence scheduling redesign?' assistant: 'I'll run the kit-knowledge-curator against prds/, my-features/, and competitor-analysis/ to find prior context.'

Use this agent when the user has created or is working on a PRD, research document, hypothesis, strategy proposal, or competitive analysis that would benefit from a critical review. Lewis should be invoked proactively after the user finishes drafting or significantly revising a document, or when the user explicitly asks for feedback on their thinking. Examples: - Example 1: user: "I just finished drafting the PRD for the new webhook retry feature" assistant: "Let me have Lewis take a look at your PRD to surface any blind spots or risks." - Example 2: user: "Here's my hypothesis: creators who use visual automations are 3x more likely to upgrade" assistant: "That's interesting. Let me have Lewis pressure-test it." - Example 3: user: "I wrote up my competitive analysis of Mailchimp's automation features" assistant: "Let me get Lewis to review it and make sure you're not missing any angles."