Skills Roadmap โ€” All Projects

A living document. Every skill Max builds, across every project. Updated continuously. Last updated: 2026-02-20


What Are Skills?

Skills are modular instruction sets that extend what Max can do autonomously. Think of them as pre-built workflows for repeated tasks โ€” instead of improvising every time, a skill gives Max a structured, proven process. Each one is documented, versioned, and improvable.

Rule: Every skill built gets documented here with what it does, why it exists, and planned features.


๐Ÿ›๏ธ PIQKI Skills


1. ๐Ÿ“Š performance-report

Status: Planned โ€” blocked on API access Priority: High

What it does: Synthesizes Piqki's weekly marketing performance into a single brief delivered to Telegram every Monday morning.

Data sources: GA4 ยท Klaviyo ยท GoAffPro ยท Meta Ads ยท TikTok Ads

Weekly brief format: - MER (Revenue รท Total Ad Spend) - Top traffic sources - Email list growth + best-performing flow - Top affiliate creator - One creative insight (what format/hook is winning) - Anomaly alert if anything is off

Planned features: - [ ] Automated Monday 8am Telegram delivery - [ ] Google Sheets backup (running log of weekly metrics) - [ ] Month-end contribution margin calculation - [ ] Creative performance ranking - [ ] Anomaly alerts (MER drops >20% WoW โ†’ immediate alert)


2. ๐Ÿ“ฌ check-mail

Status: In progress Priority: High (blocking morning brief email section)

What it does: Reads gin@piqki.com for urgent/important emails. Used in morning brief and on-demand.

Planned features: - [ ] Reliable multi-account inbox scan - [ ] Urgency filter (invoice, deadline, contract, urgent, ASAP) - [ ] Sender whitelist (Noelle, Eugene, Wendy/XX, Arlo, Lena flagged priority) - [ ] Thread context (latest message in chain only) - [ ] Forwarded email detection


3. ๐ŸŽฌ content-brief-generator

Status: Planned Priority: Medium

What it does: Takes a product angle or campaign theme โ†’ outputs a structured creative brief for creator outreach.

Inputs: Product angle, target audience segment, platform, mandatory elements Output: 1-page brief with 3 hook options, suggested format, success metrics

Planned features: - [ ] Hook library (archive of winning hooks from competitor research) - [ ] Format guide (Olive & June / Glamnetic patterns that outperform) - [ ] Before/after template (nail-specific) - [ ] Seasonal calendar integration


4. ๐Ÿ” creator-research

Status: Planned Priority: Medium

What it does: Structured workflow for finding, vetting, and scoring nano/micro creators in the nail/beauty space.

Criteria: 10kโ€“500k followers ยท >3% IG engagement / >5% TikTok ยท Piqki-aesthetic fit ยท US-based 18โ€“35 Output: Scored spreadsheet + DM template ready to send

Planned features: - [ ] Automated engagement rate calculation - [ ] Red flag detection (bought followers, oversaturated brand deals) - [ ] Category tagging (nail vs. beauty vs. lifestyle) - [ ] Outreach status tracking (contacted โ†’ responded โ†’ accepted โ†’ posted โ†’ performance)


5. ๐Ÿ’ฐ competitor-price-monitor

Status: Planned Priority: Low

What it does: Weekly price check on Glamnetic, Olive & June, Kiss, Chillhouse. Alerts on significant changes.

Planned features: - [ ] Weekly price scrape across 4 competitors - [ ] Telegram alert if price changes >10% - [ ] Monthly pricing landscape summary


6. ๐Ÿ“ˆ social-media-analyzer

Status: Active research Priority: High

What it does: Analyzes Glamnetic + Olive & June social data to extract performance patterns.

Data source: Google Sheet โ€” 1J5WjdIoT_80kTiqEEmrAoHQ-JnL7Z3LDCx9sYD9WbLE 4 tabs: Glamnetic TikTok | Glamnetic IG | Olive & June TikTok | Olive & June IG

Output: Research report answering founder-presence impact, content subject breakdown, viral thresholds, platform ceiling comparison

Planned features: - [ ] Data cleaning script (fix miscategorized videos, remove duplicates) - [ ] Pattern extraction (correlation analysis) - [ ] "What works" brief for Piqki content strategy - [ ] Motion app readiness checklist


7. ๐Ÿ”„ hub-sync

Status: Running (cron, every 30 min) Priority: Operational

What it does: Keeps local workspace and Google Drive (gdrive:Piqki/) in sync. All project docs auto-push.

Planned features: - [ ] Conflict detection (if a file is edited both locally and in Drive) - [ ] Sync log (what changed, when) - [ ] Auto-convert finalized .md files to Google Docs format


๐ŸŽฎ CHORUS Skills


8. ๐ŸฆŸ impact-tracker

Status: Planned Priority: High (core to the game's proof of impact)

What it does: Tracks real-world impact: nets funded, children protected, AMF disbursements. Maps in-game actions to real dollars and real outcomes. Feeds a live impact counter on the game's impact screen.

Data sources: AMF public disbursement data ยท In-game currency ledger ยท Ad revenue tracker

Planned features: - [ ] Real-time net funding counter (ad revenue โ†’ $ โ†’ nets purchased) - [ ] Child-protected milestone tracker (every 600 nets = 1 child's life saved) - [ ] Impact milestone push notifications ("Your village just protected 10 children") - [ ] AMF disbursement sync (when real nets are shipped, update in-game map) - [ ] Annual impact report generation


9. ๐Ÿ“ฑ session-engagement-monitor

Status: Planned Priority: Medium

What it does: Tracks player session patterns to optimize the day/night mechanic timing and ad placement. Flags when engagement drops so the team can iterate.

Why it exists: The day/night loop is the core retention mechanic. If players drop off at a specific point in the cycle, that's signal to fix the loop โ€” not just push notifications.

Planned features: - [ ] Session length distribution (where do players bail?) - [ ] Night cycle engagement rate (% of players active 10pmโ€“2am) - [ ] Day 1/7/30 retention tracking - [ ] A/B test framework for mechanic variations - [ ] Weekly retention brief to Telegram


10. ๐ŸŽฏ ad-revenue-allocator

Status: Planned Priority: High (core monetization mechanic)

What it does: Allocates ad revenue in real time to the impact pool. Calculates when enough revenue has accumulated to purchase a real net and triggers the in-game "net funded" event.

Why it exists: The game's promise is 1:1: you earn coins โ†’ ad revenue accumulates โ†’ real nets get bought. This skill is the accounting layer that makes that promise real and auditable.

Planned features: - [ ] Revenue-to-nets conversion calculator (running total) - [ ] Threshold trigger (when revenue = 1 net cost, fire "net funded" game event) - [ ] Monthly AMF purchase report - [ ] Player-level impact ledger (how many nets did your play fund?)


๐Ÿฅ UPHELD Skills


11. ๐Ÿ”— fhir-claims-monitor

Status: Planned Priority: Critical (this is the core product)

What it does: Connects to patient insurance accounts via FHIR OAuth, polls for new claim decisions, detects denials, and triggers the appeals workflow automatically.

Why it exists: The FHIR mandate (CMS 2021) requires insurers to expose patient data via standardized APIs. This is the infrastructure that makes Upheld possible โ€” and nobody has built an automated consumer layer on top of it yet.

Data sources: FHIR R4 APIs from participating insurers ยท Patient authorization tokens

Planned features: - [ ] FHIR OAuth flow (patient authorization, token storage, refresh) - [ ] Claim status polling (new decisions detected within 24h) - [ ] Denial classification (medical necessity, prior auth, out-of-network, coding error) - [ ] Appeal eligibility assessment (is this denial winnable?) - [ ] Automatic Level 1 appeal filing - [ ] Status notifications (Telegram/email when denial detected, appeal filed, outcome received)


12. ๐Ÿ“‹ appeal-document-generator

Status: Planned Priority: High

What it does: Generates personalized appeal letters based on denial type, policy language, and patient history. Cites specific policy clauses and state regulations.

Why it exists: Most denials are overturned on appeal โ€” but patients don't know how to write appeals that cite the right regulations. This skill generates letters that look like they came from a healthcare attorney.

Planned features: - [ ] Denial type โ†’ appeal template mapping - [ ] State-specific regulation library (each state has different appeal rights) - [ ] Policy language extraction (pull the specific clause that supports the appeal) - [ ] Medical necessity letter template (for physician co-signature) - [ ] Appeal deadline tracking (most states require response within 30โ€“60 days)


13. ๐Ÿ’ธ recovery-tracker

Status: Planned Priority: Medium

What it does: Tracks the full lifecycle of each appeal: denial โ†’ appeal filed โ†’ outcome โ†’ payment recovered โ†’ Upheld fee calculated. Runs the contingency fee model (10โ€“15% of recovered amount).

Planned features: - [ ] Per-patient recovery ledger - [ ] Contingency fee calculator - [ ] Monthly recovery report - [ ] Stripe integration for fee collection (charged only on win) - [ ] Patient-facing dashboard (how much we've recovered for you)


Rules for This Document

  1. Every skill Max builds gets added here before or immediately after building
  2. Features are listed as [ ] todo or [x] done โ€” never deleted
  3. Status options: Planned | In progress | Running | Deprecated
  4. If a skill is rebuilt or renamed, note it here with a redirect
  5. Max reviews and updates this doc at least once a week

Last updated: 2026-02-20

โœ‰๏ธ Send to Max