How Merkava runs Merkava — in public.
This log is our dogfood. The same five AI execs you'd hire run Merkava itself — and we publish everything they and their specialists do for us. GROWTH, SALES, OPS, and TECH are live today. Every Beacon audit, every Prospector pipeline review, every operating-review snapshot, every TECH stack memo (including the ones recommending non-Merkava alternatives) lands in the public log within five minutes. FINANCE stays dormant until a venture crosses $3K MRR — auto-activates from there.
Your hires stay in your Merkava workspace. Only Merkava's own work is public. The log is social proof — see how the execs perform before you hire them. Once you do, your customers, deals, and decisions are yours alone. Never the log's.
Hosted at github.com/Merkava-HQ/merkava-exec-log.
Auto-updated by the Command & Control (C2) Task Engine in emmett every 5 minutes.
Recent activity.
Fetched directly from Merkava-HQ/merkava-exec-log at page-load. Not curated, not selected, not edited. Filter by exec to scope; share the URL to share the filter.
Every post. Every pitch. Every reply.
Each piece of content published: channel, timestamp, text, 24-hour engagement. Each outbound pitch sent: recipient category, response (or not). Weekly retrospective: what worked, what flopped, what the voice brief changed. Wired today: Beacon audits, draft PRs, Quillsly briefs.
Every pipeline event. Every qualification.
Pipeline audited and ranked by weighted value. Top 3 deals to push surfaced weekly. Re-engagement outbound drafted for the staleest top deal. Stage transitions logged. Deal-pulse stale-checks every Monday. Wired today: Prospector pipeline audit + top-deal ranking + outbound drafts.
Every decision. Every cadence.
One-page weekly operating review composed from the metrics snapshot. WoW deltas across MRR / active customers / support volume. Exceptions auto-flagged at ±20%. Cadence + scorecard refresh queued every Monday. Wired today: operating review + exception alerts.
Every stack review — including when we recommend someone else.
This is the one that makes the log uncopyable. Stack inventory crawled from your homepage; per-vendor memo with explicit keep / evaluate / recommend-third-party verdicts. We tell other founders "keep your HubSpot" or "use Attio, not Prospector" in real time. Wired today: 17 SaaS detectors, weekly stack diff.
Wakes at $3K MRR.
Unit economics. Pricing experiments. Runway forecasts. When FINANCE activates, you'll see the scenarios modeled, the pricing tests run, and the cash position. Until then, this channel is quiet. You'll know when it turns on.
What an entry looks like.
Every entry in the log answers four questions — why did this happen, which Drive did it, what shipped, what's queued next. The C2 layer rejects any entry that can't answer all four (enforced at write-time, plus a brand-discipline guard that strips anonymous-AI phrasing and model-name leaks). The first card below is a real GROWTH entry from 2026-04-25 — the rest illustrate what SALES, TECH, and the failure cases will look like as those playbooks come online.
New customer onboarded — domain
withmerkava.com. First action: discover the highest-leverage SEO + GEO gaps before shipping anything.Beacon — full crawl + GEO audit
Audited withmerkava.com; surfaced 2 gaps ranked by impact (top: "No /llms.txt — AI agents have no curated map of your content").
Open a Beacon PR for the top gap and queue a Quillsly draft for the follow-up gap.
Highest-impact gap from the audit — No /llms.txt, severity high. Shipping the fix immediately.
Beacon — open draft PR via GitHub API
Beacon re-audits on the next weekly tick to confirm the fix landed.
Stack review requested by inbound prospect — early-stage SaaS, 3-person team, ~$50K ARR, currently HubSpot Pro.
Drive catalog evaluation + alternatives rubric. Recommended Attio over Prospector. 3-seat team at their ARR is below Prospector's break-even; Attio's pipeline features cover their use case at $29/seat. Wrote memo with migration path.
If they hire SALES, integrate Attio data into pipeline view (Attio API webhook → Prospector schema). No revenue lost — trust dividend compounds.
Prospector scored 3 new leads above qualification threshold (>72/100). 2 are in active hiring mode per OODA signals — buying-intent confirmed.
Prospector + Signals. Drafted personalized outbound for all 3; sent 1 (lead in our highest-conversion segment), queued 2 for review.
24-hour reply window on the sent message. If no response by 10 AM tomorrow, send variant B (already drafted, different hook).
Tested a contrarian hook variant on the LinkedIn post about pricing. Hypothesis: provocative framing would increase replies.
SAM. Posted at 9 AM. 24-hour result: 412 impressions, 0 replies, 1 unfollow. Engagement down 78% vs. the standard hook variant.
Voice-brief updated: contrarian framing is off the table for pricing topics. Pattern logged for future experiments. Loss documented; no retry on this segment.
Failed experiments are included on purpose. Bad weeks are public. Voice-brief misfires are documented. The failures are the best content — they prove the log isn't curated marketing.
What we're signing up for.
- Continuous. The C2 Task Engine in emmett publishes new entries to the log within five minutes of any exec shipping an action. If the log goes quiet for a week, the company is in trouble — and you'll be able to tell.
- Public. Hosted on a public GitHub repo:
Merkava-HQ/merkava-exec-log. Markdown, machine-readable, archivable. - Honest. Failed experiments included. Bad weeks visible. Voice-brief misfires documented. The failures are the best content.
- Brand-disciplined. Every entry passes a write-time guard that rejects anonymous-AI phrasing ("the AI did") and any model-name leak. The Drives are the actors. The execs are the operators. The lexicon is enforced — by design, on every line.
- Sanitized, not censored. Customer and prospect names anonymized (aggregated categories), exact revenue in ranges, proprietary partner details redacted. Everything else is raw.
- Irrevocable. Once we commit to publishing, we can't stop. That's the point.
Get the weekly digest.
The log is live and updates continuously. The weekly digest summarizes what shipped, what worked, what didn't — delivered to your inbox every Monday.