Claude Code Email Intel
Private analysis. Enter the hub password to continue.
That's not right. Try again.
📬 Private Intel Report
Claude Code Email Analysis
Emails reviewed: 1,433
Batches: 28
Trashed: 497
Kept: ~936
Confirmed paid subs: 5
Upgrade watchlist: 16
← Back to Hub
Batch Progress Summary
28 batches · 1,433 emails| Checkpoint | Total Reviewed | Total Trashed | Block Kept | Block Trashed |
|---|---|---|---|---|
| Batches 1-8 (tracked) | 115 | 24 | 91 | 24 |
| Accelerated marked pass + checkpoints 1-8 | 797 | 212 | , | 81 (older set) |
| Checkpoint 9 | 837 | 220 | 32 | 8 |
| Checkpoint 10 | 880 | 238 | 25 | 18 |
| Checkpoint 11 | 930 | 257 | 31 | 19 |
| Checkpoint 12 | 980 | 274 | 33 | 17 |
| Checkpoint 13 | 1030 | 300 | 24 | 26 |
| Checkpoint 14 | 1080 | 326 | 24 | 26 |
| Checkpoint 15 | 1121 | 343 | 24 | 17 |
| Checkpoint 16 | 1146 | 351 | 17 | 8 |
| Checkpoint 17 | 1167 | 358 | 14 | 7 |
| Checkpoint 18 | 1194 | 370 | 15 | 12 |
| Checkpoint 19 | 1217 | 379 | 14 | 9 |
| Checkpoint 20 | 1246 | 396 | 12 | 17 |
| Checkpoint 21 | 1268 | 402 | 16 | 6 |
| Checkpoint 22 | 1292 | 409 | 17 | 7 |
| Checkpoint 23 | 1318 | 418 | 17 | 9 |
| Checkpoint 24 | 1346 | 428 | 18 | 10 |
| Checkpoint 25 | 1369 | 441 | 10 | 13 |
| Checkpoint 26 | 1393 | 449 | 16 | 8 |
| Checkpoint 27 | 1416 | 463 | 9 | 14 |
| Checkpoint 28 (final) | 1433 | 473 | 7 | 10 |
Completion verification: Final Gmail query label:"Claude Code" -in:trash -label:"Claude Code Reviewed" returned 0 unreviewed messages. All 1,433 emails labeled; 497 total moved to Trash.
Confirmed Paid Subscriptions
Active as of May 24, 2026| Newsletter / Billing Name | Paid Period | Price | Relevance |
|---|---|---|---|
| Emerging AI / Opinion AI | Apr 19, 2026, Apr 19, 2027 | $70/yr | Claude app-building guide and roadmap. Strong value confirmed across multiple batches. |
| LearnAIWithMe / Gencay | Apr 28, 2026, Apr 28, 2027 | $100/yr | NotebookLM + Claude Code startup research workflow; paid 30-build series; Obsidian skill. Exceptional value. |
| Build to Launch / Jenny Ouyang | Apr 20, 2026, Apr 20, 2027 | $120/yr | Claude Skill building, MCP shipping guide, token optimization, mentor pipeline. Exceptional value. |
| GenAI Unplugged / Dheeraj | Apr 28, 2026, Apr 28, 2027 | $79.97/yr | Break-even skill, Substack Notes automation, YouTube research agent, n8n validation. Exceptional value. |
| Smart Prompts For AI | Apr 4, 2026, Apr 4, 2027 | Price not read yet | 5 Claude in Chrome Prompts slash-command issue unreviewed. High priority to assess. |
Also paid, not yet fully evaluated for Claude Code workflow value: Cash & Cache ($70/yr), AI + IQ ($60/yr), The AI Maker ($96/yr). Nate's Substack and AI Maker / Wyndo are confirmed paid and showing strong value. Smart Prompts For AI has paid-access posts visible in the Substack Paid inbox.
Upgrade Candidates Watchlist
16 candidates · price check before buying| Newsletter | Signal Strength | Known Price | Next Action |
|---|---|---|---|
| Agentic AI / Ken Huang | Leading candidate | Check price | 15-chapter harness series (permissions, compaction, MCP, orchestration). Strongest technical depth found in the sweep. |
| Daniel / Why Try AI | Strong candidate | Check price | Claude Code + Obsidian second brain, 100+ use-case swipe file, skills/MCP archive. Multiple paid-resource leads confirmed. |
| Prosper / Ilia Karelin | Strong candidate | Check price | Routines, /loop, notebooklm-py eval loop, /insights to CLAUDE.md loop. Multiple detailed free workflow issues. |
| The Digital Creator / Sharyph | Strong candidate | $39/mo or $150/yr | Skills 2.0 guide, content-research agent architecture, Substack Employee skill. Within price ceiling. |
| The AI Architect / Tyler Folkman | Good candidate | Check price | Session log harness audit loop, local multi-model router with cost logging. Real usage data. |
| Excellent AI Prompts / Lea | Good candidate | Check price | Weekend product build with CLAUDE.md + MEMORY.md, income workflow, prompt portability. Strong free preview. |
| Leadership in Change / Joel Salinas | Low-cost candidate | $49/yr | Decision-advisor skill, three downloadable marketing skills. Low-cost, within cap. |
| Standout Systems / Teodora | Good candidate | Check price | Completion-clause patterns, PreToolUse/PostToolUse safety hooks. Two strong consecutive workflow issues. |
| AI Marketing Generalist / Kieran Flanagan | Good candidate | Check price | Revenue workflows: /foundation/, compact CLAUDE.md, MCPs, Weekly Pipeline Review skill. |
| Ryan Stax / The AI Handbook | Candidate | $100/yr | Cowork skill-building method, Skill-file writing system. Compare against cheaper options first. |
| AI for Operators / Julia Kruslin | Monitor | Check price | HTML presentation workflow, .claude/agents/ogilvy-email-writer.md agent pattern. Need one more strong issue. |
| Creators AI | Monitor | Check price | Managed Agents architecture, chained skills, Motion microsite service workflow. Originality needs verification. |
| AI Blew My Mind / Daria Cupareanu | Monitor | $100/yr | Lead-magnet generation skill, AIBMM MCP, Managed-agent feedback product. Promising if Claude-specific issues continue. |
| Grow With AI / Aniket Chhetri | Monitor | Check price | Claude Code + NotebookLM research workflow reducing cost; paid tier gates the NotebookLM SKILL.md. |
| Claude Cowork / Cowork Operator | Monitor | Check price | Standing-packet/live-artifact framework, project-boundary continuity guidance. Need more evidence before upgrade. |
| Cozora | Not yet | $720/yr ($360 w/ Level Up discount) | Weekly workshops, prompt library. Named creators overlap paid subs. Use 7-day trial or $99 one-time bundle first. Annual not justified. |
Rashid / Chief Leverage Officer: YouTube content is substantive (Claude Code for business, AI employee scheduling, skill-to-plugin model). But current paid offer is $850/qtr cohort or $297/$300 standalone products. Exceeds low-cost subscription threshold. Keep free material; skip paid.
Top Tool, Skill & Repo Candidates
Highest-priority leads from all 28 batches| Candidate | Source | Why Worth Trying |
|---|---|---|
| github.com/charlie947/ai-second-brain | Charlie Hills | Public Obsidian + Claude Code second brain with /today, /ideas, /create commands. High-priority inspect. |
| notebooklm-mcp-cli and notebooklm-py | Jenny Ouyang (paid) | NotebookLM-to-Claude setup with exact install commands; cited research, logs, content repurposing, daily briefs. |
| slopcheck | ToxSec / AI and Cybersecurity | Supply-chain safety for AI agents: scan dependency files before installation; add to pre-commit or CI. |
| mattpocock/skills + browser-harness, caveman, code-review-graph, GitNexus | Ownly TEA / Code Pointer | Reusable skills, browser control, context/cost reduction, code-review assistance. |
| github.com/charlie947/social-media-skills | Charlie Hills / MarTech AI | Voice-building, newsletters, reels, thumbnails, pinned comments, scoring, hooks, analytics skills. |
| github.com/searchbrat/singleangle (Opposite-Start) | Kieran Flanagan | MIT-licensed skill for generating defensible contrarian content angles rather than generic AI writing. |
| npx agentize / github.com/stanhus/youragent | Trilogy AI Center | Agent scaffold with durable memory, BEADS task ledger, patterns library, CLAUDE.md, and OpenClaw config. |
| github.com/kenhuangus/agentctl | Ken Huang / Agentic AI | Intent-based access control across Claude Code, OpenClaw, Gemini CLI; Cedar policies, audit trails. |
| Cognee | Daily Dose of Data Science | Agent-memory combining relational provenance, vector retrieval, and graph traversal. When markdown-only stops working. |
| rtk-ai/rtk | Ownly TEA | Token-reduction repository for agent workflows. Test before accepting efficiency claims. |
| Gitleaks, TruffleHog, slopcheck, Socket, Semgrep | ToxSec | Security stack specifically for AI-authored code. Run before deploying anything agent-built. |
| rtk, OpenHarness, OpenMontage, Beads, Clauditor, repowise, autoskills | GitTrends | High-density roundup: skills, harnesses, video production, memory, token control. |
| neo4j-labs/agent-memory | Decoding AI / Paul Iusztin | Open-source graph memory with MCP tools, extraction, deduplication, retrieval. |
| High-value skills/repos mega-roundup | Daily Dose of DS (Checkpoint 12) | rowboat, superpowers, InsForge, brightdata/skills, context7, claude-mem, everything-claude-code, planning-with-files, getsentry/skills, claude-quickstarts, web-quality-skills, n8n-mcp, claude-reflect, cc-devops-skills, claude-code-skills, claude-code-plugins-plus-skills. |
| agent-efficiency-kit ($49) | Pawel Jozefiak | Low-cost agent efficiency resource within cap. Evaluate before purchase. |
| /eval-suite-wizard and /auto-research | Cash & Cache (paid) | Skill-quality loop using baselines, graders, revision logs, rollback, and cost caps. |
| YouTube research agent architecture | Dheeraj / GenAI Unplugged (paid) | Blueprint using YouTube Data API, research/note connectors, discover/audit/shorts/thumbnail modes. |
| Twelve MCP production patterns | Bilgin Ibryam / Generative Programmer | Intent tools, auth/external flows, dynamic tool loading, structured UI, server-distributed skills. |
| MemPalace commands | She Writes AI Community | /mempalace:init, :status, :help for assembling reusable voice guidance. |
| CLAUDE.md instruction-layer audit prompt | Unstuck / Sabahudin Murtic | Complete copy-paste prompt for duplicate/conflict/reactive-patch audit with cleaned-file proposal. |
| SOUL.md identity file | Prompt Guy | Nine-section structure for identity, values, boundaries, tool policy, memory, workflow, example outputs. |
| Spec-first Claude Code framework | Addy Osmani / Elevate | Living SPEC.md with gated /plan, /task, /spec, /dev-todos, /code-review, /debug, /decision workflows. |
| Omnara (mobile/web Claude Code continuation) | TechnoBizzVault | Continues Claude Code work from mobile or web with approval notifications. Worth verifying current state. |
Trashed Senders Reference
High-trash senders flagged for unsubscribe| Sender / Newsletter | Trashed | Reason Pattern | Keep? |
|---|---|---|---|
| Claude Code Architects (Skool) | 5 | Event reminders only | Mute events |
| The New Claw Times | 5 | Claude/agent industry news without reproducible workflow | Unsubscribe |
| CoolDeep AI | 6 | General tool/news commentary; practical issues were retained | Selective keep |
| Linas's Newsletter | 6 | Recaps without implementation guidance; prompting-playbook retained | Selective keep |
| The Rundown AI | 4 | General AI-news digests; some tutorials retained | Selective keep |
| World of AI | 4 | General AI/Claude product news; some instruction retained | Selective keep |
| Lenny's Newsletter | 3 | Podcast/career posts; strong workflow issues retained | Selective keep |
| Rashid / Chief Leverage Officer | 3 | Sales-only countdowns; substantive issues retained | Keep free, skip paid |
| The AI Corner | 3 | Broad founder/strategy commentary; connector ranking retained | Selective keep |
| Incredible AI | 3 | General AI news digest without Claude workflow | Unsubscribe candidate |
| Chase AI Community (Skool) | 3 | Webinar/event announcements only | Mute events |
| AI Automation Network (Skool) | 2 | Community digest without actionable method | Low signal |
| Karen / Wondering About AI | 2 | Claude behavior experiments and commentary; no executable workflow | Low signal |
| Mark Kashef / Early AI-Dopters | 3 | Generic membership promotion; polyskill issue retained | Selective keep |
Clearest unsubscribe candidates: The New Claw Times, AI Secret, World of AI (news-only issues), Byte-Sized AI, AI Insights, Incredible AI, Nextool AI, Skool event digests. These senders have produced the most repeated low-signal issues across all batches.
Unreliable for Claude Code guidance (misleading feature attributions found): AI Edge, The AI Brief, Gao Dalie.
Sender Signal Summary
Overall verdict per sender · cross-batch synthesis| Sender | Overall Signal | Best Content Found |
|---|---|---|
| Build to Launch / Jenny Ouyang | Exceptional paid | MCP shipping guide, token optimization ($1,600 bill story), mentor pipeline, Substack growth, Routines scheduler. |
| GenAI Unplugged / Dheeraj | Exceptional paid | Break-even skill, Substack Notes automation, YouTube research agent, n8n validation contracts, Channels vs OpenClaw audit. |
| LearnAIWithMe / Gencay | Strong paid | NotebookLM + Claude startup research, Obsidian problem-solver skill, competitor-intelligence agent, Substack Notes voice workflow. |
| Nate's Substack | Strong paid | Agent taxonomy framework, nate-agentic-harnesses skill, MCP/A2A protocol map, OB1/skills repo, skill design patterns. |
| Cash & Cache | Strong paid | Chained skills pipeline, eval-suite-wizard, Claude Design production guidance. Paid Notion Skill and debugging guide. |
| Emerging AI / Opinion AI | Strong paid | Claude Code masterclass (CLAUDE.md template, preflight skill, MCP), agent-stack guide, memory article, model comparison. |
| Agentic AI / Ken Huang | Leading non-sub candidate | 15-chapter harness series: typed tools, schemas, permissions, budgets, audit trails, compaction, multi-agent coordination. |
| Daniel / Why Try AI | Strong candidate | Claude Code + Obsidian second brain, 100+ use-case swipe file, .claude/commands/brainstorm.md, skills/MCP archive. |
| Prosper / Ilia Karelin | Strong candidate | Routines, /loop, notebooklm-py grounded eval, /insights loop, Agent Teams operating rules, Telegram + Channels setup. |
| Addy Osmani / Elevate | High-quality free | Spec-first framework, evidence-first code review contract, comprehension-debt safeguards, controlled coding workflow. |
| Ethan Mollick / One Useful Thing | High-quality free | Durable work patterns/risks, delegation framework, agent safety and local file warnings. Strong judgment essays. |
| Lenny / How I AI | High-quality free | Opus-build/Codex-review, accessibility workflow, Claude Code for PMs, Intercom engineering workflows, Rieseberg Cowork takeaways. |
| Matt Paige | Strong free | Granola API + CLAUDE.md recipes (meetings to apps/PRDs/ads), LLM Wiki, second-brain setup. |
| Decoding AI / Paul Iusztin | Strong free | /research_create, /research_search, /research_distill with Obsidian + Readwise + NotebookLM. /ralph-loop pattern. |
| GitTrends | Free repo discovery | Dense roundups of skills, harnesses, video production, memory, token control repos with code-level analysis. |
| Amandeep Khurana | Free high-priority | Spec-first OS with gated /spec, /design, /dev-todos, /new-feature, /code-review. Always-on agent governance model. |
| AI with Allie | Selective keep | /UE urgent-email workflow, OpenClaw use-case catalog, Agent Teams setup rules. Some high-risk use cases. |
| Grace Leung | Selective keep | Skills vs sub-agents decision guide, plugin-distribution ladder, marketing system design patterns. |
| The Rundown AI | Selective keep | Claude in Slack, Context7 guide, business tutorial leads, Excel data-cleaning workflow. Skip news-only issues. |
Key Workflow Patterns Extracted
Reusable tactics from the retained set| Pattern | What It Does | Source |
|---|---|---|
| Fixture-based regression testing | Save a brief, source set, output structure, review bar, and baseline so Cowork workflows can be rerun after product or model changes. | Claude Cowork / Cowork Operator |
| Claude + Codex adversarial review loop | Claude drafts a plan, Codex adversarially reviews it, revise, repeat before building or shipping. | Mark Kashef / Early AI-Dopters |
| 40-60% context budget + phase compaction | Clear checklist for longer coding-agent tasks: early compaction, isolated subagents, small blast radius. | Anthony Maio |
| Chained skills with context:fork | One slash command orchestrates research, curation, analysis, and formatting without polluting the main context window. | Cash & Cache (paid) |
| NotebookLM grounded skill evaluation | Build and evaluate skills from documentation rather than memory; uses notebooklm-py install + source-query workflow. | Prosper / Ilia Karelin |
| me.md + hooks + cron job second brain | Five-phase Claude Code second brain: context injection, automatic /learn capture, and deliberate knowledge growth. | Peter Yang / Ryan Wiggins |
| Token-waste audit categories | From 133,087 assistant turns: rereads, retries, dead ends, subagent overhead. Use as a cost-audit framework. | Pawel Jozefiak |
| Nine durable skill-design principles | Activation metadata, progressive disclosure, execution steps, rebuttals, deterministic scripts, scope, decay, test runs. | Bilgin Ibryam / Generative Programmer |
| /insights to CLAUDE.md improvement loop | Turns recent session friction into durable instructions. Report at ~/.claude/usage-data/report.html. | Prosper / Ilia Karelin |
| autoresearch skill improvement loop | Define yes/no evaluation criteria, batch skill improvements, test results, retain only gains. References Karpathy autoresearch. | The Digital Creator / Sharyph |
| Skill-chained second brain (Snipd + Obsidian + ElevenLabs) | Progressive-disclosure indexing, transcript capture, article-to-audio ingestion, research-to-draft chaining. | Dheeraj Sharma (paid) |
| Session log harness audit loop | Treat agent session logs as ground truth; track compaction failures, tool errors, subagent failure rates, baseline deltas. | Tyler Folkman / The AI Architect |
| Delegated email+presentation pipeline | Gmail MCP extracts messy info, creates structured spreadsheets, Cowork produces presentation deliverables. | Nitin Sharma / Ilia Karelin |
| OpenClaw three-tier hardening | Loopback binding, Tailscale/SSH, containers, LiteLLM credential brokering, Squid domain allowlists, command approvals. | AI Maker / Wyndo |
🔒 Private · annettethompson.com/hub/claude-code-email-intel · Session expires when browser closes