talgacapri
UserAI operating system for product managers. 65 Claude Code skills, 7 multi-perspective review agents, a memory system. Battle-tested in real PM work.
Categories
Indexed Skills (50)
consolidate-memory
Consolidate memory from the last 24 hours of local Claude logs (~/.claude) into memory/recent-memory.md, memory/long-term-memory.md, and memory/project-memory.md; promote durable facts and patterns from recent to long-term, and run nightly via launchd automation.
media-memory
Multimodal long-term memory. Ingest, describe, embed (Gemini Embedding 2 / gemini-embedding-001), and search any media (images, video, audio, documents) stored under /media-memory. Supports semantic similarity search plus structured metadata filtering by type, source, date range, and tag. Use when the user shares any media file, when the assistant generates any media, or when a past asset might be relevant to the current task. Triggers on: log this image, save this media, find that screenshot, do we have a recording of, search media memory, what was that file about.
connect-mcps
Connect MCPs for real-time tool integration
activation-analysis
Analyze user activation using Setup → Aha → Habit framework. Identifies activation bottlenecks.
ai-prioritize
Prioritize AI features using Confidence-Gated Prioritization. Replaces RICE/WSJF/MoSCoW for AI products with four AI-native lenses.
competitor-analysis
Deep competitive analysis + ongoing monitoring. Checks user research for competitor mentions, sales notes, existing analysis. Integrates with retention-analysis and user-research-synthesis.
decision-doc
Document important product decisions. Creates decision logs with rationale, alternatives, and trade-offs.
define-north-star
Identify and validate your North Star Metric. Aligns product strategy with key business metric.
expansion-strategy
Upsell, cross-sell, and account growth tactics. Framework for revenue expansion.
experiment-decision
Decide when to A/B test vs just ship. Framework for experiment planning and prioritization.
experiment-metrics
STEDII framework for selecting trustworthy experiment metrics. Ensures metric validity and reliability.
feature-metrics
Define success metrics using the STEDII framework for trustworthy experiment metrics.
feature-results
Post-launch analysis and results documentation. Document what shipped and what we learned.
impact-sizing
Quantify feature value with driver trees, confidence levels, and the 4-step sizing framework.
interview-guide
Create JTBD-based interview guides for user research. Structured questions for discovery interviews.
meeting-cleanup
Batch process multiple meetings from a single day. Consolidates action items and insights across meetings.
meeting-notes
Transform meeting transcripts into structured action items, decisions, and key insights. Processes raw notes, voice memos, or recordings.
metrics-framework
Set up leading vs lagging indicators for product decisions. Framework for metric selection and tracking.
morning-brief
Scan industry news, summarize important emails, check today's calendar, and send a 60-second brief to Slack DM. Runs automatically via GitHub Action at 6 AM GST or on-demand. Triggers on: morning brief, daily brief, news brief, what did I miss, catch me up, brief me.
pg-pressure-test
Pressure-test any startup idea, product bet, or new feature with the brutal honesty of a Paul Graham YC application review. Six named passes — Pressure Test the Idea (fatal flaws), Validate the Real Problem (vitamin vs painkiller), Map the Real Competition (current behavior + indirect), Find the First 10 Customers (manual traction plan), Build the MVP in 2 Weeks (riskiest assumption only), and the Brutal Verdict (strong / weak / pivot). Use when the user says "pressure test this idea", "is this a real startup", "would PG fund this", "kill or keep this idea", "is this a vitamin or a painkiller", "what's the riskiest assumption", "find me my first ten customers", "scope an MVP", "brutal feedback on this concept", or pastes a startup idea, feature concept, or one-line pitch. Triggers on `/pg-pressure-test`, `/pressure-test`, `/paul-graham`, `/brutal-verdict`, or any explicit request for PG-style critique.
prd-draft
Create a modern, AI-era PRD for features and initiatives. Guides through clarifying questions, generates draft, and offers multi-agent review.
prioritize
Classify PM tasks using LNO Framework (Leverage/Neutral/Overhead) to focus on high-impact work.
prototype-feedback
Build → review → iterate prototype workflow. Structured feedback collection and iteration.
prototype
Advanced prototyping (Artifacts/Figma/Lovable/v0/Bolt)
ralph-wiggum
Devil's advocate PRD/document reviewer with humor and sharp critique
research-scout
Find new or contradictory external signals (web search, Reddit, Hacker News, Quora) and stage only validated learnings in memory/long-term-memory.md. Use when the user asks for new research, fresh strategies/tools announcements, or contradiction checks against existing docs.
retention-analysis
Cohort analysis and retention optimization framework. Identifies retention drivers and churn factors.
slack-message
Draft team communications for Slack. Creates clear, actionable messages for different contexts.
daily-plan
Generate PM daily plan with context
create-tickets
Create tickets via Linear/Jira MCP or generate formatted ticket text
design-backlog-writer
Create Design Epics and Design Tasks from PRDs or plans, following a JTBD-driven design backlog structure. Optimised for designer flexibility — defines WHAT data/components are needed, never dictates HOW screens should be laid out. Use when asked to "create design tasks", "write design backlog", "design epic", "break down for design", "design deliverables from this PRD", or any request to create design-oriented backlog items separate from engineering stories.
frontend-design
Create distinctive, production-grade frontend interfaces in your product's brand voice. Use this skill when building web components, pages, prototypes, dashboards, HTML/CSS layouts, or any web UI. Generates polished, intentional code that feels genuinely yours, not generic AI-generated output. Always reads your design system tokens before producing any output. Triggers on: "build a page", "create a component", "make a dashboard", "prototype this", "HTML for", "design a screen", "landing page", or any request to style or beautify a web interface.
generate-ai-prototype
Generate v0.dev, Lovable, or Bolt.new prompts for AI-powered prototyping
interview-feedback
Post-PM-interview debrief and continuous improvement for job search
interview-prep
Pre-interview preparation for PM job interviews (Product Sense, Execution, Behavioral)
jefferson-fisher
Coach the user through any conversation using Jefferson Fisher's Next Conversation methodology — the Three Pillars (Control, Confidence, Connection), the Hidden Conversation, named tools (the 2-1-6 Breath, the Pause, Smart Reset Questions, the Two-Question Defuser, Set the Frame), and a power phrase library for boundaries, bad news, saying no, rude people, interruptions, criticism, and conflict. Use when the user says "help me say this", "what should I say to", "how do I respond when", "I have a tough conversation", "argue less", "communicate with confidence", "set a boundary", "deliver bad news", "say no", "stand up for myself", "stop being talked over", or pastes a message they need to send/answer in a hard conversation. Triggers on "/jefferson-fisher", "/next-conversation", or any direct request for Fisher-style communication coaching.
journey-map
Create user journey maps and customer journey maps (dual mode)
mece
Apply MECE (Mutually Exclusive, Collectively Exhaustive) thinking to break down any problem, structure options, or organize a framework. Use when structuring a problem, building a driver tree, listing options in a decision doc, defining user segments, identifying root causes, or organizing any set of categories that must not overlap and must cover everything. Triggers on "break this down", "structure this problem", "MECE check", "am I missing anything", "are these exhaustive", "do these overlap", or any request to decompose a problem, decision, or analysis into clean categories.
meeting-agenda
Create structured meeting agendas for effective collaboration
meeting-feedback
Post-meeting effectiveness feedback and continuous improvement
napkin-sketch
ASCII wireframes + browser capture for design matching
roadmap-view
Generate and update a visual, interactive product roadmap from a JSON data file. Produces a branded HTML swimlane view with quarters, sprints, and feature blocks — ready to share with stakeholders. Triggers on: "show roadmap", "update roadmap", "generate roadmap", "roadmap view", "visual roadmap", "sprint roadmap", "share the roadmap", "roadmap HTML", or any request to see or update the product roadmap. Also auto-prompts for roadmap placement when a new PRD or feature concept is created.
sketchnote
Generate a branded visual overview of any feature from a PRD. Produces a copy-paste-ready prompt for Claude, ChatGPT, or any image-capable AI. Output is sized for Notion embeds and Slack shares (1080px width). Sits between /prd-draft and /generate-ai-prototype in the workflow.
code-first-draft
Initial feature implementation
launch-checklist
Comprehensive product launch planning
prd-review-panel
Multi-agent PRD review (7 perspectives)
design-sprint
Run a structured 5-day design sprint from problem framing to prototype testing. Use at the start of a project for ideation, solution selection, rapid prototyping, and concept validation.
hooked-ux
Diagnose retention and engagement issues by mapping product behavior to the Hook Model (trigger, action, reward, investment). Use when users are not coming back, engagement drops, or the team wants to improve habit loops.
ios-hig-design
Design and review iOS interfaces against Apple's Human Interface Guidelines, including safe areas, navigation, typography, accessibility, and native interaction patterns. Use for iOS UI design, SwiftUI screens, and HIG compliance reviews.
refactoring-ui
Audit interface quality using Refactoring UI principles for hierarchy, spacing, color, and depth. Use when the UI feels off, when the user asks to fix visual hierarchy, refine spacing, improve color palette, or polish visual design before ship.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.