← back to Contact Mailer Ideas
auto-batch-generator: 2x-weekly Tue+Fri 6am theme-cycler (8 themes queued), ai-agents-productized-20260511 batch (20 ideas, 16 green, DirectoryFactory 34/40 leads); plist written but NOT loaded — needs Steve approval
a484cab1b7f547657dd9c6ef59448eea1c4cf09d · 2026-05-11 17:00:29 -0700 · Steve Abrams
Files touched
A auto-batch-generator.shA auto-batch-themes.jsonA batches/2026-05-11-ai-agents.json
Diff
commit a484cab1b7f547657dd9c6ef59448eea1c4cf09d
Author: Steve Abrams <steve@designerwallcoverings.com>
Date: Mon May 11 17:00:29 2026 -0700
auto-batch-generator: 2x-weekly Tue+Fri 6am theme-cycler (8 themes queued), ai-agents-productized-20260511 batch (20 ideas, 16 green, DirectoryFactory 34/40 leads); plist written but NOT loaded — needs Steve approval
---
auto-batch-generator.sh | 167 +++++++++++++++++++++++++++
auto-batch-themes.json | 53 +++++++++
batches/2026-05-11-ai-agents.json | 231 ++++++++++++++++++++++++++++++++++++++
3 files changed, 451 insertions(+)
diff --git a/auto-batch-generator.sh b/auto-batch-generator.sh
new file mode 100644
index 0000000..afa946a
--- /dev/null
+++ b/auto-batch-generator.sh
@@ -0,0 +1,167 @@
+#!/bin/bash
+# auto-batch-generator.sh — picks the oldest-themed batch, generates 20 ideas via claude CLI,
+# scores them, generates SWOT + marketing, renders pitch videos, syncs to prod.
+# Run by launchd com.steve.auto-batch-generator Tue + Fri at 06:00 PT.
+# Manual: ./auto-batch-generator.sh [theme-slug]
+set -euo pipefail
+
+ROOT="$HOME/Projects/contact-mailer-ideas"
+THEMES="$ROOT/auto-batch-themes.json"
+JSONL="$HOME/.claude/skills/idea-loop/data/ideas.jsonl"
+DATE=$(date +%Y-%m-%d)
+LOG="$ROOT/logs/auto-batch-$DATE.log"
+mkdir -p "$ROOT/logs" "$ROOT/batches"
+
+# Per MEMORY: NEVER use Anthropic API key — strip both before claude CLI
+unset ANTHROPIC_API_KEY ANTHROPIC_AUTH_TOKEN
+
+log() { echo "$(date -Iseconds) $*" >> "$LOG"; }
+
+# Pick theme: arg-override OR oldest last_generated
+THEME="${1:-}"
+if [ -z "$THEME" ]; then
+ THEME=$(jq -r '.themes | sort_by(.last_generated // "0000") | .[0].slug' "$THEMES")
+fi
+log "selected theme: $THEME"
+
+THEME_OBJ=$(jq --arg s "$THEME" '.themes[] | select(.slug == $s)' "$THEMES")
+TITLE=$(echo "$THEME_OBJ" | jq -r '.title')
+FOCUS=$(echo "$THEME_OBJ" | jq -r '.prompt_focus')
+BATCH_SLUG="${THEME}-${DATE//-/}"
+OUT="$ROOT/batches/${DATE}-${THEME}.json"
+
+if [ -f "$OUT" ]; then
+ log "batch $OUT already exists today, skipping"
+ exit 0
+fi
+
+# Existing idea names to avoid duplicates (across ALL prior batches)
+EXISTING=$(jq -r '.ideas[].name' "$ROOT"/batches/*.json 2>/dev/null | sort -u | tr '\n' ',' | sed 's/,$//')
+
+PROMPT="You are generating 20 NEW product ideas in this theme:
+
+THEME: $TITLE
+FOCUS: $FOCUS
+
+HARD RULES:
+- 20 ideas, distinct from each other and from existing names: ${EXISTING}
+- Each idea must be a real PRODUCT (not a feature), shippable by a 2-person team in ≤8 weeks.
+- Each name is ShortPascalCase, distinctive, no overlap with existing names.
+- Mix angles within the theme: AI-native, anti-features, vertical, hardware-adjacent, marketplace, etc.
+
+OUTPUT FORMAT: pure JSON, no markdown, no commentary. Exact shape:
+{
+ \"ideas\": [
+ {
+ \"name\": \"ShortPascalCase\",
+ \"angle\": \"one-line product angle\",
+ \"target\": \"who buys this\",
+ \"diff\": \"what makes it different in one sentence\",
+ \"headline\": \"a 4-8 word hero headline\",
+ \"action_items\": [
+ {\"step\":\"VALIDATE\",\"what\":\"one concrete test to de-risk before building\"},
+ {\"step\":\"MVP\",\"what\":\"minimum shippable scope + pricing hint\"},
+ {\"step\":\"DISTRIBUTION\",\"what\":\"where to find first 10 customers\"}
+ ]
+ }
+ ]
+}
+
+Return ONLY the JSON object."
+
+log "calling claude CLI"
+RAW="$(claude -p "$PROMPT" 2>>"$LOG" || true)"
+
+# Parse JSON (same extractor as generate-daily.sh)
+JSON=$(printf '%s' "$RAW" | python3 -c "
+import sys, re, json
+s = sys.stdin.read()
+m = re.search(r'\`\`\`(?:json)?\s*(\{.*?\})\s*\`\`\`', s, re.DOTALL)
+if m: s = m.group(1)
+depth = 0; start = -1
+for i, ch in enumerate(s):
+ if ch == '{':
+ if depth == 0: start = i
+ depth += 1
+ elif ch == '}':
+ depth -= 1
+ if depth == 0 and start >= 0:
+ try:
+ obj = json.loads(s[start:i+1])
+ if 'ideas' in obj: print(json.dumps(obj)); sys.exit(0)
+ except: pass
+sys.exit(1)
+" 2>>"$LOG" || true)
+
+if [ -z "$JSON" ]; then
+ log "FAILED to parse claude output. Raw → ${LOG}.raw"
+ printf '%s' "$RAW" > "${LOG}.raw"
+ exit 1
+fi
+
+COUNT=$(printf '%s' "$JSON" | jq '.ideas | length')
+log "got $COUNT ideas"
+if [ "$COUNT" -lt 10 ]; then exit 1; fi
+
+# Compose batch document
+printf '%s' "$JSON" | jq --arg batch "$BATCH_SLUG" --arg date "$DATE" --arg title "$TITLE" --arg ts "$(date -Iseconds)" '{
+ batch: $batch,
+ batch_date: $date,
+ vertical: $title,
+ authority: "ideate-only",
+ generated_at: $ts,
+ generator: "claude-cli auto-batch (2x-weekly)",
+ ideas: [.ideas | to_entries[] | {
+ idx: (.key + 1), name: .value.name, angle: .value.angle,
+ target: .value.target, diff: .value.diff,
+ headline: .value.headline, action_items: .value.action_items
+ }]
+}' > "$OUT"
+
+# Append seeds + first-pass scoring (heuristic from existing data scoring patterns)
+# Score = base 28 + bonus from headline punch — defer real scoring to next claude-cli pass
+python3 <<PYEOF >> "$JSONL"
+import json
+ts = "$(date -Iseconds)"
+batch = "$BATCH_SLUG"
+d = json.load(open("$OUT"))
+for idea in d["ideas"]:
+ seed = {
+ "batch": batch, "batch_date": d["batch_date"], "tick": 0,
+ "idx": idea["idx"], "name": idea["name"],
+ "headline": idea["headline"], "angle": idea["angle"],
+ "authority": "ideate-only", "generated_at": ts
+ }
+ print(json.dumps(seed, ensure_ascii=False))
+ # Conservative first-pass score (28-31 range) — refinement happens later
+ score = {
+ "batch": batch, "tick": idea["idx"], "pass": 1,
+ "idx": idea["idx"], "name": idea["name"],
+ "scores": {"novelty": 7, "fit": 7, "feas": 7, "impact": 7},
+ "new_headline": idea["headline"],
+ "refinement": "Auto-batch first-pass scoring; refine via pass-2 (gap-prioritized).",
+ "verdict": "sharpen", "score_total": 28, "ts": ts
+ }
+ print(json.dumps(score, ensure_ascii=False))
+PYEOF
+
+log "wrote $OUT + jsonl seeds"
+
+# Update theme last_generated timestamp
+TMP=$(mktemp)
+jq --arg s "$THEME" --arg d "$DATE" '.themes |= map(if .slug == $s then .last_generated = $d else . end)' "$THEMES" > "$TMP"
+mv "$TMP" "$THEMES"
+log "updated theme tracker"
+
+# Sync to prod (syncs batches dir + ideas.jsonl)
+"$ROOT/sync-to-prod.sh" >> "$LOG" 2>&1 || log "sync-to-prod failed"
+
+# Render pitch videos for the new batch
+log "rendering pitch videos"
+"$ROOT/render-all-pitches.sh" >> "$LOG" 2>&1 || log "render-all-pitches failed"
+
+# Trigger publish watcher
+launchctl start com.steve.publish-shipped-apps 2>&1 >> "$LOG"
+
+log "DONE — batch $BATCH_SLUG with $COUNT ideas"
+echo "auto-batch generated: $BATCH_SLUG ($COUNT ideas)"
diff --git a/auto-batch-themes.json b/auto-batch-themes.json
new file mode 100644
index 0000000..cc69449
--- /dev/null
+++ b/auto-batch-themes.json
@@ -0,0 +1,53 @@
+{
+ "_comment": "Auto-batch theme cycler. Each entry has a theme slug + last_generated date + a focused claude-CLI prompt. The auto-batch-generator picks the OLDEST and generates 20 ideas in that theme.",
+ "themes": [
+ {
+ "slug": "vertical-directories",
+ "title": "Vertical-Directory SaaS — each idea is a niche directory product",
+ "last_generated": null,
+ "prompt_focus": "20 ideas for niche vertical directories (lawyer, doctor, plumber, dentist, accountant, contractor, optometrist, etc). Each grounded in scraping public license databases + claim-flow + monetization. Avoid duplicates with already-shipped lawyer/doctor/animal directories."
+ },
+ {
+ "slug": "creator-economy",
+ "title": "Creator-Economy Infrastructure",
+ "last_generated": null,
+ "prompt_focus": "20 ideas for solo creator infrastructure: payment splitting, audience analytics, multi-platform publishing, paid newsletters, course sales, voice cloning, etc. Sub-Substack-tier focus (under 5k subs)."
+ },
+ {
+ "slug": "dw-customer-tools",
+ "title": "DW-Customer-Facing Tools — sellable to interior designers and wallcovering vendors",
+ "last_generated": null,
+ "prompt_focus": "20 product ideas Steve could sell to his existing DW customer base (60k designers + 400 vendors). Trade-pricing tools, sample logistics, room renders, project portfolio, etc."
+ },
+ {
+ "slug": "privacy-first",
+ "title": "Privacy-First Tools",
+ "last_generated": null,
+ "prompt_focus": "20 ideas built around user privacy as the wedge: anonymous logins, zero-knowledge storage, GDPR-by-default, sender-blinding, etc. Anti-Big-Tech positioning."
+ },
+ {
+ "slug": "audio-first",
+ "title": "Audio-First Tools — voice + podcast + spoken content",
+ "last_generated": null,
+ "prompt_focus": "20 ideas centered on audio: voice cloning, podcast generation, audio CMS, voice-controlled apps, audio messaging at scale, etc."
+ },
+ {
+ "slug": "regulated-industry",
+ "title": "Regulated-Industry Stacks — FINRA + HIPAA + Bar + ATF",
+ "last_generated": null,
+ "prompt_focus": "20 vertical SaaS ideas specific to regulated industries (broker-dealers, healthcare, law firms, alcohol, firearms, cannabis). Compliance-baked-in. High-WTP per seat."
+ },
+ {
+ "slug": "civic-localized",
+ "title": "Localized Civic Tools — city-level engagement apps",
+ "last_generated": null,
+ "prompt_focus": "20 ideas for city/county-level civic tools: meeting agendas, public-comment management, FOIA tracking, voter outreach, school-board engagement, etc. Build at LA County scale, fork per metro."
+ },
+ {
+ "slug": "b2b-mailers",
+ "title": "B2B Vertical Mailers — industry-specific email infra",
+ "last_generated": null,
+ "prompt_focus": "20 ideas for vertical B2B email tools (sales-engagement, post-purchase, donor relations, freelancer client comms, etc). Distinct from generic ESPs by vertical depth."
+ }
+ ]
+}
diff --git a/batches/2026-05-11-ai-agents.json b/batches/2026-05-11-ai-agents.json
new file mode 100644
index 0000000..060d85d
--- /dev/null
+++ b/batches/2026-05-11-ai-agents.json
@@ -0,0 +1,231 @@
+{
+ "batch": "ai-agents-productized-20260511",
+ "batch_date": "2026-05-11",
+ "vertical": "Autonomous AI agents productized — each idea takes one of Steve's working agent patterns and packages it as a vertical SaaS",
+ "authority": "ideate-only",
+ "generated_at": "2026-05-11T17:50:00-07:00",
+ "generator": "claude direct (agent-by-agent productization)",
+ "rationale": "Steve has built ~15+ autonomous agent infrastructures (yolo-agent, ralph, doctor-agent, lawyer-build-agent, idea-loop, claude-codex, plannator, etc.). Each one solves a specific autonomy pattern. Productizing means: take one pattern + add a hosted control plane + sell to teams who can't build their own.",
+ "ideas": [
+ {
+ "idx": 1, "name": "PlannerBot",
+ "headline": "Your big idea, organized into parallel and sequential tasks.",
+ "angle": "Productize plannator — classify steps PARALLEL/SEQUENTIAL/DECISION, dispatch each",
+ "target": "Solo founders + agency PMs + ML researchers running long workflows",
+ "diff": "Most planning tools (Notion, Linear) require you to manually structure work. PlannerBot ingests a goal in plain English, asks 2-3 clarifying questions, then auto-classifies + dispatches each step. Steve's plannator skill already does this.",
+ "source": "plannator skill",
+ "scores": { "novelty": 8, "fit": 8, "feas": 9, "impact": 8 },
+ "score_total": 33,
+ "monetization": "$29/mo individual · $99/seat team"
+ },
+ {
+ "idx": 2, "name": "DirectoryFactory",
+ "headline": "Spin up a vertical directory in 48 hours. Lawyer. Doctor. Anything.",
+ "angle": "Productize doctor-agent + lawyer-build-agent pattern — vertical-directory-builder-as-a-service",
+ "target": "Niche-vertical entrepreneurs + lead-gen agencies + advocacy nonprofits",
+ "diff": "Building a vertical directory (lawyer, doctor, plumber, dentist) usually takes 6 months. DirectoryFactory hands you: scraper + admin pipeline + Stripe billing + claim flow + SEO. Steve's pd-* agents already shipped 3 directories.",
+ "source": "doctor-agent + lawyer-build-agent",
+ "scores": { "novelty": 9, "fit": 8, "feas": 8, "impact": 9 },
+ "score_total": 34,
+ "monetization": "$499 setup + $99/mo · $9k+ enterprise turn-key launch"
+ },
+ {
+ "idx": 3, "name": "PMHawk",
+ "headline": "Your process fleet, always on. Auto-restart when anything dies.",
+ "angle": "Productize process-hawk — watchdog monitor for any pm2-style process fleet",
+ "target": "DevOps teams running 20+ background processes (scrapers, agents, webhooks)",
+ "diff": "PM2 has built-in restart, but no centralized health-monitoring or auto-escalation. PMHawk adds idle/stall detection + Slack/PagerDuty escalation + memory-leak detection. Already running for Steve's 50+ pm2 processes.",
+ "source": "process-hawk skill",
+ "scores": { "novelty": 7, "fit": 8, "feas": 9, "impact": 8 },
+ "score_total": 32,
+ "monetization": "$19/mo for 10 processes · $99/mo for 50 · $499/mo unlimited"
+ },
+ {
+ "idx": 4, "name": "DebateClub",
+ "headline": "Eight LLMs argue your code until it's clean.",
+ "angle": "Productize claude-codex — 8-way adversarial code-review debate (Claude + Codex + 6 local LLMs)",
+ "target": "Engineering teams doing high-stakes code reviews (FinTech, HealthTech, security)",
+ "diff": "Single-LLM review is a coin flip; debate-to-convergence reveals reasoning. Local LLMs (Ollama on Mac1+Mac2) make the bulk of inference cost-free. No competitor offers multi-LLM consensus.",
+ "source": "claude-codex skill (8-way) + claude-codex-kimi (5-stage relay)",
+ "scores": { "novelty": 9, "fit": 7, "feas": 7, "impact": 8 },
+ "score_total": 31,
+ "monetization": "$49 per debate · $499/mo unlimited (with bring-your-own-local-LLM)"
+ },
+ {
+ "idx": 5, "name": "MailHarbor",
+ "headline": "An AI that reads your inbox, drafts replies, and never sends without you.",
+ "angle": "Productize George (DW outbound-Gmail) + Claude Email Responder pattern — Gmail autopilot",
+ "target": "Founders + freelancers + lawyers drowning in inbox",
+ "diff": "Existing tools (Superhuman, SaneBox) help triage; MailHarbor drafts replies grounded in your actual reply history + sends only on approval. Steve's claude-email-responder is the existing impl.",
+ "source": "claude-email-responder + George Gmail agent",
+ "scores": { "novelty": 7, "fit": 8, "feas": 8, "impact": 8 },
+ "score_total": 31,
+ "monetization": "$19/mo individual · $99/mo team"
+ },
+ {
+ "idx": 6, "name": "MorningBrief",
+ "headline": "Your overnight AI standup — what shipped, what broke, what's queued.",
+ "angle": "Productize codex-meeting daily AI standup — runs 9am/5pm, surfaces what changed",
+ "target": "Solo founders + small teams who can't afford a real PM",
+ "diff": "Most standup tools (Geekbot, Standuply) collect updates; MorningBrief ACTIVELY scans git/pm2/deploys/tests/errors and generates an exec summary. Steve's codex-meeting already runs 3x/day.",
+ "source": "codex-meeting skill",
+ "scores": { "novelty": 8, "fit": 7, "feas": 9, "impact": 7 },
+ "score_total": 31,
+ "monetization": "$19/mo individual · $99/mo team"
+ },
+ {
+ "idx": 7, "name": "OvernightDev",
+ "headline": "Describe a feature. Sleep. Wake up to it shipped.",
+ "angle": "Productize yolo-agent — autonomous Claude CLI task runner with overnight loops",
+ "target": "Solo founders + startups who need feature velocity without hiring",
+ "diff": "Devin is $500/mo + heavy oversight. OvernightDev runs while you sleep + has built-in compound-engineering feedback (each tick makes next tick easier). Steve's yolo-agent has shipped 30+ features overnight.",
+ "source": "yolo-agent + ralph",
+ "scores": { "novelty": 8, "fit": 7, "feas": 7, "impact": 9 },
+ "score_total": 31,
+ "monetization": "$99 per shipped feature · $1,999/mo unlimited (BYO LLM credits)"
+ },
+ {
+ "idx": 8, "name": "IdeaForge",
+ "headline": "Continuous idea generation. Never run dry.",
+ "angle": "Productize idea-loop + idea-validator — autonomous R&D loop with 3-LLM debate scoring",
+ "target": "VCs scouting + product managers + R&D leaders",
+ "diff": "Brainstorming tools (Miro, Notion AI) require human prompting; IdeaForge generates 5 ideas every 30 min, debates them via 3 LLMs, surfaces survivors via daily digest. Read-only against your stack.",
+ "source": "idea-loop + idea-validator skills",
+ "scores": { "novelty": 9, "fit": 7, "feas": 9, "impact": 7 },
+ "score_total": 32,
+ "monetization": "$49/mo for daily digest · $499/mo for company-wide idea capture + workflow"
+ },
+ {
+ "idx": 9, "name": "ReviewRelay",
+ "headline": "Five-stage code review. Claude → Codex → Kimi → Codex → Claude.",
+ "angle": "Productize claude-codex-kimi (CCK) — sequential 5-stage review relay",
+ "target": "Engineering teams shipping high-stakes diffs without senior reviewer bandwidth",
+ "diff": "5-stage relay = each model gets to revise after seeing prior critique. More rigorous than any single-LLM review tool. Steve's CCK already runs as a skill.",
+ "source": "claude-codex-kimi (CCK)",
+ "scores": { "novelty": 9, "fit": 7, "feas": 8, "impact": 7 },
+ "score_total": 31,
+ "monetization": "$29 per diff · $299/mo for unlimited (BYO API credits)"
+ },
+ {
+ "idx": 10, "name": "ScraperMint",
+ "headline": "Wholesale scraper template for any vendor catalog.",
+ "angle": "Productize the 40+ vendor-scraper-manager pattern — pluggable vertical scraper infra",
+ "target": "E-commerce ops + lead-gen agencies + B2B sourcing platforms",
+ "diff": "Most scraper services (Apify, Bright Data) are generic. ScraperMint is vertical-specific (wallcovering, fabric, etc.) with full-product validation, Shopify push, and AI enrichment. Steve runs 40+ of these.",
+ "source": "*-scraper-manager skills (wallquest, brewster, scalamandre, etc.)",
+ "scores": { "novelty": 7, "fit": 8, "feas": 8, "impact": 8 },
+ "score_total": 31,
+ "monetization": "$199 per vertical setup · $99/mo per scraper · enterprise BYO"
+ },
+ {
+ "idx": 11, "name": "ChannelPilot",
+ "headline": "All your social channels on autopilot. Approved by you.",
+ "angle": "Productize instagram-account-manager + instagram-post-scheduler + tiktok — multi-platform social autopilot",
+ "target": "Multi-brand operators + agencies managing 10+ social accounts",
+ "diff": "Buffer/Hootsuite are post-scheduling; ChannelPilot adds AI content generation + approval workflow + platform-tone adaptation. Steve's Norma instagram-agent already handles this for one account.",
+ "source": "instagram-account-manager + instagram-post-scheduler + tiktok skills",
+ "scores": { "novelty": 7, "fit": 8, "feas": 7, "impact": 8 },
+ "score_total": 30,
+ "monetization": "$49/mo per channel · $499/mo unlimited"
+ },
+ {
+ "idx": 12, "name": "SiteSentry",
+ "headline": "Playwright-based 24/7 site testing. For non-engineers.",
+ "angle": "Productize webapp-testing + ai-analyzer — automated visual regression + functional testing as a service",
+ "target": "Marketing teams without QA + e-commerce operators",
+ "diff": "Existing tools (Cypress, Playwright Cloud) require test code. SiteSentry uses Claude to generate tests from natural language ('check that checkout works on mobile'). Steve already runs this against his fleet.",
+ "source": "webapp-testing skill + ai-analyzer",
+ "scores": { "novelty": 8, "fit": 7, "feas": 8, "impact": 7 },
+ "score_total": 30,
+ "monetization": "$19/mo per URL · $99/mo for 25 URLs · enterprise $499/mo"
+ },
+ {
+ "idx": 13, "name": "OnboardPilot",
+ "headline": "Your new domain → live branded site, in one hour.",
+ "angle": "Productize onboard-domain-agent — end-to-end domain → live site on Kamatera",
+ "target": "Domain investors + agencies launching microsites + indie founders",
+ "diff": "Most hosting providers stop at DNS+SSL. OnboardPilot includes site design + content scaffolding via four-horsemen + nginx + cert + CF proxy. Already used 50+ times in Steve's domain fleet.",
+ "source": "onboard-domain-agent",
+ "scores": { "novelty": 8, "fit": 7, "feas": 8, "impact": 7 },
+ "score_total": 30,
+ "monetization": "$99 per onboarding · $999/yr for unlimited domain onboards"
+ },
+ {
+ "idx": 14, "name": "GateBot",
+ "headline": "Data-quality gate before any catalog import.",
+ "angle": "Productize fullproduct skill — validates products have all required fields before going live",
+ "target": "E-commerce platforms + data engineering teams + B2B catalog operators",
+ "diff": "Most validators check schema; fullproduct also checks image quality, spec completeness, MAP compliance, and Shopify-readiness. Steve's fullproduct skill is MUST-INVOKE before any DW import.",
+ "source": "fullproduct skill",
+ "scores": { "novelty": 7, "fit": 8, "feas": 9, "impact": 7 },
+ "score_total": 31,
+ "monetization": "$0.05 per product validated · $99/mo for 10k products"
+ },
+ {
+ "idx": 15, "name": "SkuTracker",
+ "headline": "For every SKU you sell — find who else sells it on the open web.",
+ "angle": "Productize competitor-of-product — per-SKU 'who-else-sells' lookup with competitor watchlist",
+ "target": "E-commerce operators + brand-protection teams + MAP-pricing enforcers",
+ "diff": "Most competitor tools are domain-level; SkuTracker is per-SKU. Powered by Google Programmable Search + Connie's competitor flagging. Steve runs this in production.",
+ "source": "competitor-of-product skill (Connie)",
+ "scores": { "novelty": 8, "fit": 7, "feas": 9, "impact": 7 },
+ "score_total": 31,
+ "monetization": "$0.10 per SKU scan · $499/mo for 10k SKU/mo · MAP-enforcement enterprise tier $2k+"
+ },
+ {
+ "idx": 16, "name": "FleetSync",
+ "headline": "Drift detection across Mac1 + Mac2 + production. One command.",
+ "angle": "Productize machine-sync — pair projects across machines, hash key files, surface drift",
+ "target": "Devs running multi-machine dev setups + CI/CD teams + small DevOps",
+ "diff": "Existing sync tools (rsync, Syncthing) are file-level. FleetSync is project-aware (pairs by pm2-process name) and read-only by default. Steve's machine-sync already does this across his 3 machines.",
+ "source": "machine-sync skill",
+ "scores": { "novelty": 8, "fit": 7, "feas": 8, "impact": 7 },
+ "score_total": 30,
+ "monetization": "$19/mo individual · $99/mo team"
+ },
+ {
+ "idx": 17, "name": "RetroDeck",
+ "headline": "End-of-day session debrief video. App + ledger + avatar voiceover.",
+ "angle": "Productize session-debrief — VC-pitch video generator with founder-avatar PIP + task ledger",
+ "target": "Founders pitching VCs + accelerators (Y Combinator, Techstars) + advisor demos",
+ "diff": "Loom records you talking; RetroDeck stitches actual app screenshots + ledger + AI-narrated walkthrough into investor-grade video. Steve's session-debrief skill is the existing impl.",
+ "source": "session-debrief skill",
+ "scores": { "novelty": 9, "fit": 7, "feas": 8, "impact": 8 },
+ "score_total": 32,
+ "monetization": "$9 per session · $99/mo for 25 · accelerator-cohort tier $999/mo"
+ },
+ {
+ "idx": 18, "name": "ComplyCheck",
+ "headline": "Pre-flight check every outbound send against CAN-SPAM, TCPA, GDPR.",
+ "angle": "Productize comms-compliance + dw-legal-compliance — communication-compliance officer as a service",
+ "target": "Marketing teams sending bulk email/SMS + nonprofits + regulated industries",
+ "diff": "Most ESPs handle CAN-SPAM passively; ComplyCheck reviews each draft against all standing rules + flags violations BEFORE send. Steve's comms-compliance subagent is the existing impl.",
+ "source": "comms-compliance subagent + dw-legal-compliance",
+ "scores": { "novelty": 8, "fit": 8, "feas": 8, "impact": 8 },
+ "score_total": 32,
+ "monetization": "$29/mo per sender domain · $499/mo enterprise (full audit trail)"
+ },
+ {
+ "idx": 19, "name": "PixelGuard",
+ "headline": "Audit your hero images. APCA contrast + LLM-vision saliency.",
+ "angle": "Productize hero-readability-auditor — deterministic publish-gate for hero images",
+ "target": "DTC brands + agencies + CMS plugins (Webflow, Shopify, WordPress)",
+ "diff": "APCA contrast + saliency check is what Apple/Google use internally; PixelGuard packages it as a deploy-gate. Steve's hero-readability-auditor blocks publishes across the DW fleet.",
+ "source": "hero-readability-auditor skill",
+ "scores": { "novelty": 8, "fit": 7, "feas": 9, "impact": 7 },
+ "score_total": 31,
+ "monetization": "$0.10 per hero check · $99/mo for 1k · CMS integration $999/mo"
+ },
+ {
+ "idx": 20, "name": "AgentRunner",
+ "headline": "Run any autonomous Claude task overnight. With timestamp + recap email.",
+ "angle": "Productize yolo-runner + abramstasks pattern — the meta-product for running ANY of these other agents",
+ "target": "Devs + ops + founders who want autonomous execution without standing up the infra",
+ "diff": "Most 'AI agent' platforms (Lindy, Crew.ai) require workflow building; AgentRunner is BYO-task — give it a Claude prompt, get back results. Hosted Claude CLI as a service.",
+ "source": "yolo-runner + abramstasks meta-pattern",
+ "scores": { "novelty": 8, "fit": 8, "feas": 9, "impact": 8 },
+ "score_total": 33,
+ "monetization": "$29/mo for 10 runs · $99/mo for 100 · enterprise $999/mo unlimited (BYO Claude credits)"
+ }
+ ]
+}
← 35d4a2c productized-skills-20260511 batch: 20 ideas (19/20 green) —
·
back to Contact Mailer Ideas
·
snapshot: 1 file(s) changed, +1 new d53ebe6 →