Hermes Repo — Complete Map
This repository is the working brain of 10xers Labs: every research doc, strategy, prompt, transcript, and operational script generated by the Hermes agent since June 2026. It’s big (~700 files, 90+ knowledge docs) — this page is the map.
Quick links: Web site (this, rendered) · GitHub repo · Graph view (right sidebar)
The Two Kinds of Content
| Kind | What | Where | On the site? | On your Mac (Syncthing)? |
|---|---|---|---|---|
| Knowledge docs | Research, strategy, guides, prompts, analyses — what you read | Most top-level folders | ✅ Yes | ✅ Yes |
| Operational | Scripts, configs, credentials, automation — what runs things | scripts/, hermes-config/, profiles/, adspower-dashboard/, hermes-clone-kit/ | ❌ No (may hold secrets) | ❌ No (VPS-only) |
| Chat transcripts | Hermes session logs — chronological record, not knowledge | hermes-chat-transcripts/ | ❌ No | ❌ No |
| Video transcripts | Raw material for analysis (YouTube etc.) | video-transcripts/ | ✅ Yes (indexed) | ✅ Yes |
Knowledge Folders (what you read)
| Folder | What’s inside | Start with |
|---|---|---|
guides/ | Setup guides, deep-dives, this map | obsidian-syncthing-setup, knowledge-base-visualization-comparison |
growreach-reddit-strategy/ | The Reddit marketing playbook: strategy guide, subreddit tiers, swarm PRD/flow, synthesized research | reddit-marketing-strategy-guide |
reddit-rpa-prompts/ | The comment-system prompts (v2→v5) that drive RPA commenting | v5 (current) |
reddit-ban-forensics/ | Why accounts got banned — BotBouncer source analysis, detection research, data checklist | BOTBOUNCER-SOURCE-CODE-ANALYSIS |
competitor-analyses/ | Linqin AI competitive intelligence (product, homepage, study) | linqin-ai-competitor-analysis |
cro-analysis/ | CRO reports from PostHog data (agent reports, deep-dives, PostHog audit) | growreach-cro-deep-dive-report |
adspower-kb/ | AdsPower browser-automation knowledge base (API, CLI, MCP) | 00-index |
linkedin-lead-gen-guide/ | LinkedIn lead generation framework | — |
brand-strategy/ | GrowReach split-screen brand strategy | — |
rockethub-analysis/ | RocketHub / AppSumo lifetime-deal pricing research | — |
video-transcripts/ | Raw YouTube transcripts (Gojiberry, SynQ-with-Saad) — analysis material | — |
notebooklm-custom-prompts/ | NotebookLM custom mining prompts | — |
Operational Folders (what runs things — VPS-only, not on site)
| Folder | What | Why it’s excluded |
|---|---|---|
scripts/ | autolink.py, frontmatter.py, sync-daemon.sh, fetch scripts | May reference credentials/proxies |
hermes-config/ | agent-hooks (task-rules.md, enforcement), profiles | Internal agent config |
profiles/ | Hermes profile configs (reddit-marketer, cro, etc.) | Internal |
adspower-dashboard/ | AdsPower automation: swarm engine, dashboard, proxy mgmt | Contains proxy/IP/password plans |
hermes-clone-kit/ | Windows EC2 sync pack + profile templates | Internal tooling |
web/ | Quartz site source (this site’s engine) | Build tooling |
skills/, skill-bundles/ | Hermes skill copies | Mirrors of installed skills |
Raw / Chronological
| Folder | What |
|---|---|
hermes-chat-transcripts/ | 166 Hermes session transcripts (auto-synced every 30 min by cron) — searchable history, not knowledge docs |
sarvam-experiment/, reddit-wiki-fetch/, split-screen-experiment/, voice-overlay/, autonomous-ai-agents/ | Experiments and scratch — kept for reference |
How It All Connects
┌─────────────────────────────────────────────┐
│ GitHub repo (this repo) │
│ single source of truth, git-versioned │
└──────────────┬──────────────────────────────┘
│
┌────────────────────┼────────────────────┬─────────┐
▼ ▼ ▼ ▼
Quartz → Cloudflare Obsidian (Mac) Hermes agent Syncthing
Pages (site, graph view + reads files (Mac/Android
graph + search, backlinks, for context, get knowledge
shareable links) editing writes docs docs only)
- Site (this page’s home): https://hermesvps.pages.dev — rebuilt + deployed on every push, $0/month, unlimited deploys
- Obsidian on Mac:
~/hermes-reposynced via Syncthing — graph view over the same files, edits auto-commit to GitHub within ~2 min - Agent: reads these files for context; every new doc ships with frontmatter +
## Relatedwikilinks (Rule 6)
FAQ
Why isn’t scripts/ or profiles/ on the site? They contain operational secrets (proxies, API keys, IP plans). The site publishes knowledge only.
Why aren’t chat transcripts on the site? They’re chronological records (166 files), not knowledge — they’d flood the graph with isolated nodes. Search them via GitHub or session_search.
Why are video transcripts included? They’re raw material for analysis — Gojiberry and SynQ-with-Saad research draws on them.
Why did transcripts/ become two folders? (2026-08-14) — chat logs moved to hermes-chat-transcripts/ (excluded from site), video/analysis material moved to video-transcripts/ (included). The transcript-sync cron now writes to hermes-chat-transcripts/.
How do I find something? Site search (top-right) · GitHub search · Obsidian search on Mac · just ask the agent.
Last updated: 2026-08-14. This map is maintained alongside the repo — if a folder is missing, it’s likely new or operational.