jetthoughts · pftg · Aug 8, 2026 · Aug 8, 2026 · Aug 8, 2026 · Aug 8, 2026
diff --git a/.okf/log.md b/.okf/log.md
@@ -927,3 +927,50 @@ HTML on five posts, three of them already live (kamal-2-multi-server,
 ai-code-ownership-accountability, claude-code-xp-team-workflow). It is internal
 review metadata from the writer brief. Removed sitewide; writers should report
 cadence in the handback, never in the file.
+
+## 2026-08-08 — Passive keyword monitor purged from the sourcing machinery
+
+**Decision (Paul)**: remove F5Bot outright — "it's not working as you wanted."
+The tool was already marked retired on 2026-08-08, but the retirement was
+half-done: its doc survived as a "kept for the keyword derivation" stub, and two
+separate boards still carried *"paste the v2 keywords into F5Bot"* as a
+load-bearing 2-minute ask on Paul's desk (`operation-runbook.md` Sprint-3 header
+and `20.09` §P0). A dead blocker parked on the CEO's desk is worse than no
+blocker: it reads as the cheapest available unblock, so it displaces the real
+one.
+
+**Durable rule — two structural tests any sourcing tool must pass.** Neither is
+fixable by better keywords, and both killed this one:
+
+1. **Active, not passive.** A monitor only knows what arrived after it was
+   switched on, so it can never backfill a window. Card #29 needs rows verified
+   inside ≤30 days *on demand*; a passive feed cannot satisfy that by
+   construction.
+2. **Ranked retrieval, not literal substring matching.** Founders don't type our
+   phrasing. The v1 composed phrases (`vibe coded broke`) matched **0 of the 25
+   posts we had already found by hand**.
+
+Do not re-propose the class. Judge paid alternatives (RedReach, Buska, Syften,
+CommunityTracker) against those two tests *before* spending.
+
+**Keyword derivation preserved and re-verified**, moved into
+`prospects/p7-search-sweep.md` §3 with counts re-run across
+`voice-of-customer.md` + `cold-prospect-list.md`: `non-technical` 20×,
+`replit` 16×, `lovable` 8×, `cursor` 5×, `vibe coded` 2× (vs the gerund
+`vibe coding` 10×, rejected — it floods with commentary, not distress).
+
+**Finding surfaced by the move**: `dev shop` has **zero** corpus occurrences
+(`dev shop`/`devshop` 0, `agency` 1). It carries the Trigger-3
+ownership/hostage lane, and its only prior coverage argument *was* passive
+monitoring. So our second-strongest trigger now rides an unevidenced query term
+in a lane grooming already proved unfindable through open search. A zero return
+on it is a **lane gap to escalate**, not a quiet zero. `bolt.new`/`base44` are
+likewise unevidenced tool-name guesses — drop after two empty sweeps.
+
+**Correction to the board while here**: `#29` was described as agent-runnable.
+It is not — the replacement sweep ran on 2026-08-08 and returned zero rows
+because every thread-open hit `EGRESS_BLOCKED` and Reddit rejected the tool's
+user agent. Qualification requires a timestamp read from the *opened* thread, so
+**no sourcing method can produce a lead until thread-open access is restored**
+(`chrome-devtools` + egress to indiehackers.com / reddit.com). That is now the
+real P0 in both boards, in place of the keyword ask.
diff --git a/.okf/workflows/outbound-sprint.md b/.okf/workflows/outbound-sprint.md
@@ -6,6 +6,7 @@ tags: [outbound, sales, browser-agent, kanban, sourcing]
 generated:
   by: process:okf-migrate
   at: 2026-07-26T00:00:00Z
+updated: 2026-08-08T00:00:00Z
 resource: docs/projects/2607-vibe-code-rescue/
 sources:
   - resource: "/workflows/review-swarm.md"
@@ -21,8 +22,8 @@ sources:
    for Paul's per-message approval, then send, then log. Nothing sends on
    silence.
 2. **Monitor** - each morning `rescue-sprint/reply-monitor-prompt.md`
-   sweeps IH notifications, LinkedIn, Reddit, and Gmail (`neetocal`,
-   `from:f5bot.com`), classifies replies, drafts responses from the
+   sweeps IH notifications, LinkedIn, Reddit (once batch 2 is live), and
+   Gmail (`neetocal`), classifies replies, drafts responses from the
    objection bank (marked DRAFT - Paul sends), and updates the ledger.
 3. **Measure** - `rescue-sprint/pipeline.md` is the single ledger (one row
    per touch + Friday tally). The A0 kill-criteria (~20 touches / ~10
@@ -85,10 +86,16 @@ sources:
 
 # Gotchas
 
-- **F5Bot sends nothing until the account email is verified** - the verify
-  email hid for 3 days (2026-07-21 to 07-24) because...
-- ...mail to `pftg.sof@gmail.com` is auto-labeled + archived in that
+- **Passive keyword monitors do not work for this sprint** (retired
+  2026-08-08 after a 2-week trial). Two structural flaws, neither tunable:
+  they match literal substrings, so composed founder phrasing never fires
+  (0 hits against 25 posts we found by hand), and they are passive, so they
+  can never backfill the ≤30-day lead window on demand. Sourcing uses
+  date-filtered active search instead
+  (`rescue-sprint/prospects/p7-search-sweep.md`). Do not re-propose the
+  class - judge any paid alternative against those two tests first.
+- **Mail to `pftg.sof@gmail.com`** is auto-labeled + archived in that
   mailbox and auto-forwarded to Paul's main inbox. Monitor by search query
-  (`from:f5bot.com`, `neetocal`), never by scanning that inbox.
+  (`neetocal`), never by scanning that inbox.
 - Statusline/parallel sessions contend on `.git/index.lock` - transient;
   re-check before force-removing.
diff --git a/docs/business/operating-system.md b/docs/business/operating-system.md
@@ -1,110 +1,85 @@
 # JetThoughts Operating System (lean)
 
-**Scope**: this is the **company** operating system - how JetThoughts runs across every opportunity, not just the current one. It is the durable cadence (accountability chart, quarterly OKR, rocks, weekly scorecard, issues, pipeline). The **active bet it currently drives** is the [Vibe Code Rescue validation project](../projects/2607-vibe-code-rescue/) (portfolio bet #1); when the next opportunity becomes active, the same OS runs it with a new OKR/rocks.
-**Company layer**: [`index.md`](index.md) (front door) · [`vision-mission.md`](vision-mission.md) (who we are) · [`opportunity-portfolio.md`](opportunity-portfolio.md) (the bets, and how we validate/kill/scale them).
-**Purpose**: the smallest cadence that reliably ships the current company objective - right now, **1 Vibe Code Rescue client by Nov 30, 2026** (the active portfolio bet).
-**Owner**: Paul Keen | **Created**: 2026-07-21 | **Review**: weekly
-**Model**: EOS-lite (accountability chart + quarterly OKR + weekly scorecard + rocks + issues), sized for a founder + white-label delivery + AI execution.
+**Scope**: the **company** operating system - how JetThoughts runs across every opportunity, not just the current one. **Company layer**: [`index.md`](index.md) (front door) · [`vision-mission.md`](vision-mission.md) (who we are, still DRAFT) · [`opportunity-portfolio.md`](opportunity-portfolio.md) (the bets).
+**Purpose**: the smallest cadence that reliably ships the current objective - **1 Vibe Code Rescue client by Nov 30, 2026**.
+**Owner**: Paul Keen | **Created**: 2026-07-21 | **Review**: weekly, in §1 below - no separate ritual
+**State**: this file does NOT duplicate the runbook's fresh-session state. For "what do I do right now," open [`operation-runbook.md`](../projects/2607-vibe-code-rescue/operation-runbook.md) → START HERE. This file is the numbers layer over it.
 
 ---
 
-## 0. Current state & where to start (fresh session)
+## 1. This week
 
-**As of 2026-07-21**: foundation is set. A0 assumptions register ✓, G0 white-label partner ✓ (locked per CEO), offer/kit/booking drafts ✓, T1 trigger taxonomy ✓, T2 sourcing pipeline ✓.
-**Active sprint**: Rescue Demand Sprint — populate the prospect list (agent) so Paul can send + book calls.
-**Next agent work**: `T4/T5` (build the list) → `T6` dedupe → `T7` openers; `T8` discovery kit in parallel.
-**To execute**: open [`operation-runbook.md`](../projects/2607-vibe-code-rescue/operation-runbook.md) → **▶ START HERE** block → take the next Ready task. That runbook is the shared state across sessions; this file is the weekly cadence over it.
-**Paul's desk (human-only)**: booking page (~5 min), confirm price band (✅ done: $2,500 triage / $7,500 rescue / $10,000 foundation reset), review + send the outreach, take the calls, close. Optional: Gmail consent (warm lane), Apollo $79/mo + Hunter key (paid cold funnel).
+Source of truth: [`rescue-sprint/pipeline.md`](../projects/2607-vibe-code-rescue/rescue-sprint/pipeline.md) §Weekly tally. This table is that tally's numbers, not a second set - update the sheet first, this row second, same sitting.
+
+| week of | touches sent | replies | calls booked | audits sold | rescues signed | posts shipped |
+|---|---|---|---|---|---|---|
+| 2026-08-08 | 0 | 0 | 0 (target 8-12 by Nov 30) | 0 (target 3-5) | 0 (target 1) | 2 (Aug, target ~6/mo) |
+
+**Verdict: 🔴 red.** Not "zero calls" - **zero touches have ever been sent.** Batch-1 has been HOLD since 2026-07-26 (stale rows); the Sprint-3 re-source ran 2026-08-08 and is now BLOCKED-ON-TOOLING (`operation-runbook.md` §Card #29). 18 days since Sprint 3 kicked off (2026-07-26), the bottleneck hasn't moved past sourcing.
+**Kill-check**: the C1 kill-criteria ("2 weeks of warm outreach yields 0 booked calls → pause, re-open A+C") has **not fired** - its precondition is 2 weeks of *actual* outreach, and none has gone out yet. Don't read that as safe: it means we don't even have the data the criteria needs. See Issue 2.
+**Where sessions went this week**: mostly 2605 (course wave completion, W1-W5) and 2510 (content re-plan), plus this session's 2607 sourcing-tool diagnosis. That allocation is why a distribution failure sat unflagged for 18 days - nobody was in the pipeline sheet.
 
 ---
 
-## 1. Accountability chart (who owns what)
+## 2. Accountability chart
 
 | Seat | Owner | Owns |
 |---|---|---|
-| **CEO / Sales / Trust layer** | Paul | Discovery calls, offer, pricing, the founder relationship, fractional-CTO oversight |
-| **Delivery** | White-label partner | The rescue build (Rails), under JT brand, to JT quality bar |
-| **Marketing / Content / Ops engine** | Claude Code (+ Paul review) | Content, SEO, landing, lead magnets, LinkedIn drafts, pipeline tracking, this OS |
-
-One person can hold several seats - but every seat has exactly one accountable owner.
+| **CEO / Sales / Trust layer** | Paul | Discovery calls, offer, pricing, founder relationship |
+| **Delivery** | White-label partner | The rescue build (Rails), under JT brand |
+| **Marketing / Content / Ops** | Claude Code (+ Paul review) | Content, SEO, landing, pipeline tracking, this OS |
 
-## 2. Quarterly OKR (Q3->Q4 2026)
+## 3. OKR (Q3→Q4 2026) + kill-check
 
 **Objective**: Land the first Vibe Code Rescue client and prove the machine.
 
-| Key Result | Target | Status |
+| KR | Target | Status |
 |---|---|---|
-| KR1 - Offer + white-label partner live | Rescue Audit priced, partner signed, landing CTA live | 🟢 partner signed ✓; offer priced ✓ ($2,500 / $7,500 / $10,000); landing deferred |
-| KR2 - Discovery calls booked | 8-12 by Nov 15 | 🔲 |
-| KR3 - Audits delivered | 3-5 free Rescue Audits (45-min call + 48h scorecard — the funnel entry; paid tiers follow) | 🔲 |
-| KR4 - Rescue signed | **1 engagement by Nov 30** | 🔲 |
+| KR1 - Offer + partner live | Priced, partner signed, landing CTA live | 🟢 partner ✓, priced ✓ ($2,500/$7,500/$10,000), booking link live ✓; full landing page still Blocked (card B1) |
+| KR2 - Discovery calls booked | 8-12 by Nov 30 | 🔴 0. (Mid-Nov is an internal trend checkpoint, not a separate deadline - both KRs share one date now.) |
+| KR3 - Audits delivered | 3-5 free Rescue Audits | 🔴 0 - blocked behind KR2 |
+| KR4 - Rescue signed | 1 by Nov 30 | 🔴 0 - blocked behind KR3 |
 
-## 3. Rocks (90-day priorities = roadmap milestones)
+**Kill-criteria (from the assumptions register, evaluated here weekly - this is the "runs automatically via OS-WEEKLY" promise, made real):** C1 warm-channel test - 0 calls after 2 weeks of *actual sent* outreach → pause, re-open ICP/channel. Currently **untestable**, see §1. A0 ICP test - if ≥half of early interest is technical or pre-launch founders, re-open the ICP vote. Not yet evaluable - zero interest of any kind so far.
 
-1. **Offer + partner locked** (Aug) - the machine can't sell without these.
-2. **Landing = booking machine** (Aug) - offer, proof, one CTA.
-3. **Demand-gen live on all channels** (Aug->Sep) - LinkedIn 3-4/wk, rescue SEO cluster, paid pilot.
-4. **First audit -> first signing** (Oct).
-
-## 4. Weekly scorecard (leading metrics, review every week)
-
-| Metric | Owner | Target/wk |
-|---|---|---|
-| LinkedIn posts published | Paul/Claude | 3-4 |
-| LinkedIn qualified replies/DMs | Paul | >=3 |
-| New email subs from magnets | Claude | trend up |
-| Blog posts shipped (rescue/ICP-E) | Claude | 2-3 |
-| **Discovery calls booked** | Paul | trend to 8-12 by mid-Nov |
-| Rescue Audits sold | Paul | cumulative -> 3-5 |
-| Rescues signed | Paul | -> 1 |
+## 4. Rocks (open only)
 
-Green = hit target, Red = missed. Any red 2 weeks running becomes an Issue (section 6).
+1. **Unblock sourcing** (now) - card #29 needs `chrome-devtools` + egress to at least one of indiehackers.com / reddit.com. Nothing else on the board matters until this moves.
+2. **Landing page** (Aug) - card B1, Blocked on nothing but capacity; booking link already live standalone.
+3. **First send → first call** (Aug-Sep) - batch-1/2/3, gated entirely on Rock 1.
+4. **First audit → first signing** (Oct) - not started, gated on Rock 3.
 
-## 5. Meeting cadence
+*Closed*: offer + partner locked (Jul 21). *Cut* (20.09, 2026-08-07): the paid-pilot rock - budget only matters once organic proves a reply signal, and the bet currently forbids a content sprint.
 
-> **Async-first rule**: meetings below are the exception, not the medium. All knowledge transfer defaults to written artifacts per `.agents/skills/async-first-communication/SKILL.md`; every meeting's outcome is written back into this file (scorecard/OKR/issues) same-day.
-
-- **Weekly ops review (~30 min)**: scorecard (5 min) -> rock progress (5 min) -> pipeline
-  walk (10 min) -> IDS the top issues (10 min). Update the scorecard + OKR statuses in this file.
-- **Monthly strategy check (~45 min)**: is the wedge still right? adjust roadmap; refresh
-  market scan if signals shift.
-- **Per-deal**: discovery call -> audit -> proposal, tracked in the pipeline (section 7).
-
-## 6. Issues list (IDS: Identify, Discuss, Solve)
-
-Rolling list of the top blockers. Solve the highest-leverage one each week; don't let it grow.
+## 5. Issues (IDS - top 5, solve the highest-leverage one each week)
 
 | # | Issue | Owner | Status |
 |---|---|---|---|
-| 1 | White-label delivery partner (runbook G0) | Paul | ✅ Resolved 2026-07-21 (locked per CEO). Standing validation: confirm a paid trial rescue hits the quality bar + a fallback partner is named before the first real signing. |
-| 2 | Prospect list populated — 25 deduped rows merged in P8 | Claude | ✅ Resolved 2026-07-22 (P8 complete) |
-| 3 | Price band confirmed — A2 set at $2,500 triage / $7,500 rescue / $10,000 foundation reset | Paul | ✅ Resolved 2026-07-22 (A2 confirmed) |
+| 1 | Sourcing BLOCKED-ON-TOOLING - #29's sweep can't open any thread to verify a timestamp | Infra/Paul | 🔴 Open since 2026-08-08 - see runbook §Card #29 |
+| 2 | Kill-criteria untestable, not "not fired" - 18 days, zero touches sent, so the C1 test has never actually run | Claude/Paul | 🔴 Open - ratify: is this a distribution defect (re-run once unblocked) or does it call the bet itself into question? |
+| 3 | Joy Adamson override - only survivor of batch-1, 5mo old, still publicly unanswered | Paul | 🟡 Open - 1-min decision, unresolved since ~2026-07-26 |
+| 4 | Pricing vs. thesis tension - JT undercuts at $7,500 against $25-55K competitors one week after publishing "cheap developers are expensive," which argues against its own thesis to a twice-shy founder | Paul | 🟡 Open - content flagged it (20.09 §10), not a content fix |
+| 5 | vision-mission.md still stamped DRAFT, 18 days on, while other docs cite it as settled | Paul | 🟡 Open - 1-line decision: ship it or say what's wrong |
 
-## 7. Pipeline (simple funnel, no CRM needed yet)
+*Resolved, moved to changelog*: white-label partner (Jul 21) · prospect list populated, P8 (Jul 22) · price band (Jul 22).
 
-Track each lead as a row: **Source -> Discovery call -> Audit sold -> Proposal -> Signed**.
-Stages: `Lead` -> `Call booked` -> `Audit sold` -> `Proposal out` -> `Won` / `Lost (reason)`.
-Keep it in a single sheet (or a section here) until volume justifies a real CRM.
-**Leads originate from** the populated target list in [`rescue-sprint/warm-intro-referral-kit.md`](../projects/2607-vibe-code-rescue/rescue-sprint/warm-intro-referral-kit.md) (agent-built via runbook T4-T7); once Paul sends, each contacted row becomes a pipeline `Lead`.
-
-## 8. Tooling map
+## 6. Tooling map
 
 | Function | Tool |
 |---|---|
-| Site / blog | Hugo (`bin/hugo-build`), content in `content/blog/` |
-| Content ops | Blog pipeline (`docs/workflows/blog-pipeline.md`), voice guide 90.11 |
-| LinkedIn | LinkedIn pipeline (`docs/workflows/linkedin-post-pipeline.md`) + ICP validation plan |
-| Discovery scheduling | NeetoCal (per outreach template) |
-| Pipeline tracking | Sheet / this doc section 7 until CRM is justified |
-| OKR spine | `lark-okr` skill for structuring quarterly OKRs |
-| Weekly OS loop | `eos-lite` skill (`Skill(eos-lite) weekly`) — runs scorecard/OKR/rocks/issues/pipeline over this doc |
-| Step-by-step execution | [`operation-runbook.md`](../projects/2607-vibe-code-rescue/operation-runbook.md) — ordered task cards + Incremental agent backlog; **▶ START HERE** for any fresh session |
-| Foundational premises | [`rescue-sprint/assumptions-register.md`](../projects/2607-vibe-code-rescue/rescue-sprint/assumptions-register.md) — ICP/trigger/channel/pre-validation, voted, with kill-criteria (runbook rule 7) |
-| Sprint artifacts | `rescue-sprint/` — offer, warm-intro kit, booking spec, trigger taxonomy, sourcing pipeline (see runbook START HERE artifact index) |
-
-## 9. Definition of "operating" (are we actually running the system?)
-
-- Scorecard updated weekly, with real numbers.
-- Every rock has a next action and an owner.
-- The Issues list is short because we solve the top one each week.
-- Discovery-calls-booked moves up week over week from Sept.
+| Site / blog | Hugo (`bin/hugo-build`), `content/blog/` |
+| Content ops | `docs/workflows/blog-pipeline.md`, voice guide 90.11 |
+| LinkedIn | `docs/workflows/linkedin-post-pipeline.md` + ICP validation plan |
+| Discovery scheduling | NeetoCal |
+| Pipeline tracking | [`rescue-sprint/pipeline.md`](../projects/2607-vibe-code-rescue/rescue-sprint/pipeline.md) - the only copy of these numbers |
+| OKR/weekly-loop skills | `lark-okr`, `eos-lite` (`Skill(eos-lite) weekly`) |
+| Step-by-step execution | [`operation-runbook.md`](../projects/2607-vibe-code-rescue/operation-runbook.md) - START HERE for any fresh session |
+| Foundational premises | [`rescue-sprint/assumptions-register.md`](../projects/2607-vibe-code-rescue/rescue-sprint/assumptions-register.md) - kill-criteria live here |
+
+## Changelog (resolved issues + closed rocks move here, newest first)
+
+| Date | Change |
+|---|---|
+| 2026-08-08 | Rewrote this file: 110→~90 lines, bound §1 to `pipeline.md` instead of duplicating a second scorecard, fixed the KR2 Nov-15/Nov-30 date conflict, replaced the 3 stale "all-resolved" Issues with the real ones, cut the paid-pilot rock (superseded by 20.09), deleted the stale §0 snapshot in favor of linking the runbook. |
+| 2026-07-22 | Prospect list populated (P8, 25 rows) · price band confirmed ($2,500/$7,500/$10,000) |
+| 2026-07-21 | White-label partner locked · offer/kit/booking drafts done |