DecisionNerd

Use the imessage-analysis MCP server to answer questions about the user's iMessage history.

Compare your messaging activity between two time periods — volume, contacts, and trends side by side.

Analyse your group chat activity — most active groups, top talkers per group, and activity over time.

Check whether imessage-analysis is installed and the MCP server is registered, and fix anything that's missing.

Show the current status of the iMessage analysis dataset — last sync time, total messages, and whether it's up to date.

Summarise your most recent message exchanges — who you've been talking to, what topics came up, and the overall tone of recent activity.

Sync your iMessage history — builds the dataset on first run, updates incrementally after that. Use this to make sure your data is current before analysing.

Compare your messaging patterns with two or more contacts side by side — volume, frequency, reactions, and trajectory.

Show the balance sheet for the modelled periods — assets, liabilities, and equity.

Show the cash flow statement for the modelled periods — operating, investing, and financing activities.

Extract financial data from QuickBooks Online or SEC EDGAR and build the WACCY dataset.

Show the income statement (P&L) for the modelled periods — revenue, expenses, and net income.

Build a three-statement financial model (income statement, balance sheet, cash flow) from the extracted dataset.

Execute SQL against the WACCY financial dataset to answer specific questions about the data.

Check whether waccy is installed and the MCP server is registered, and fix anything that's missing.

Show the current status of the WACCY dataset — last extraction time, record count, and whether it's up to date.

Deploy a streamlit-private project to a managed host (Railway first; Render/Fly.io later) and return the private, authenticated URL. Use when the user wants to "deploy my Streamlit app", "ship it", "put it online privately", or "deploy to Railway".

List and act on pending access requests for a privately deployed Streamlit app — approve (add the user to the organization) or reject them via streamlit-private. Use when the user (an admin) wants to "approve access requests", "see who's requested access", or "let someone in".

Reconfigure an existing streamlit-private project — switch the auth provider (e.g. Clerk → WorkOS) or hosting provider (e.g. Railway → Render), or regenerate deployment assets, while preserving the user's application code. Use when the user wants to "change auth provider", "switch hosting", or "regenerate my streamlit-private config".

Initialize a Streamlit app for private deployment with streamlit-private — scaffold a new app or wrap an existing Streamlit repo, adding a gateway, deployment assets, and a manifest without touching application code. Use when the user wants to "make my Streamlit app private", "set up streamlit-private", or "add auth to my Streamlit app".

Invite a user to a privately deployed Streamlit app's organization via streamlit-private, so they become a member and gain access. Use when the user (an admin) wants to "invite someone", "add a user", or "give my colleague access" to the app.

Diagnose a failed streamlit-private deployment or broken private access — missing provider secrets, proxy/WebSocket issues, auth/membership problems, or a misconfigured manifest. Use when the user says "my deploy failed", "the app won't load", "sign-in is broken", or "the page is blank / keeps reconnecting".

Fills scaffolded DocSlime documents by interviewing the user and removing inline LLM guidance. Use when writing product, discovery, requirements, design, architecture, testing, publishing, observability, or resolving placeholders.

Scaffolds and tailors the DocSlime docs starting template with docslime init. Use when setting up project docs, initializing docs in a repo, or safely adding missing default docs.

Reviews a DocSlime docs tree for KISS violations, bloat, generic AI slop, weak requirements, overbuilt architecture, stale contradictions, and non-actionable design or testing docs. Use when tightening docs or applying best-practice cleanup.

Creates and fills the next-numbered DocSlime Architecture Decision Record. Use when recording a significant product or technical decision, documenting why an option was chosen, or adding an ADR.

Verifies and installs the DocSlime CLI. Use when setting up DocSlime, after installing the skill pack with npx skills, or when a docslime command is missing.

Emotion-aware product design: audit how a product, flow, or surface feels; map emotional journeys; calibrate tone, trust, delight, and friction; write emotion briefs for Impeccable; hand off to DocSlime. Use when the user wants feeling-first product design, emotional UX review, tone/microcopy feel, peak-end moments, trust and anxiety reduction, empty/error/success state emotion, persona emotional lenses, or `/productfeeling` commands. Complements Impeccable (craft) and DocSlime (docs) — does not replace them. Works in chat or with project `docs/` (DocSlime) and optional `.productfeeling/` persistence.

Review and upgrade one Product Feelings Handbook section at a time: research the method, add high-quality further reading, rewrite for clarity and ethics, and leave an Agent skill TODO stub. Use when the user asks to review, upgrade, research, or rewrite a handbook strategy, TTP, or concept; mentions SECTION_REVIEW_PROMPT; or works section-by-section on the handbook/ tree.