Communications tab (search/query roadmap items 1 & 2):
- Fix the investor dropdown: the facet only listed grid investors, so it
came back empty whenever email matched a classic contact or org domain
(no grid id — the common case). It now mirrors the email list, resolving
each link to a typed identity (fund:/org:/contact:/addr:) with precedence
grid -> org -> contact -> address; investor_id accepts the typed key
(bare id = fund: for back-compat) and an unknown prefix matches nothing.
- Add a date-range filter and a click-to-expand full-body view
(GET /api/email/detail, admin, soft-delete-gated; body_text only, never
raw remote HTML).
- Add a "Search content" mode: GET /api/email/search wraps the ingest
hybrid_search over the Qdrant email index (doc_type=email), hydrated and
soft-delete-filtered against SQLite (canonical), 503 if Spark/Qdrant down.
Daily digest:
- Settings -> Admin builds a digest over a chosen window (last 24h or since
a date) as an in-app preview before sending (POST /api/admin/digest/preview),
so the local-Spark summarizer can be verified on demand even on a quiet day.
Manual send uses the same window; neither advances the daily cursor, so a
preview never suppresses the scheduled digest.
Code-only, migrations no-op. 22/22 backend tests, render-smoke pass.
React/ReactDOM/Babel were loaded from the unpkg CDN at runtime — react@18
and react-dom@18 weren't even exact-pinned, and none had SRI. A CDN swap (or
react auto-resolving a new 18.x) could blank the whole app with no change on
our side: exactly the v78/v79 blank-screen class. It also made the self-hosted
box depend on outbound internet to render.
Vendor the three libs into frontend/assets/vendor/ (React 18.3.1, ReactDOM
18.3.1, @babel/standalone 7.29.7) and load them same-origin with sha384
integrity attributes. They now ship inside the s9pk (Dockerfile already COPYs
frontend/; server.py serves /assets/* with the path-containment check), so a
CDN can never swap prod deps again and no outbound fetch is needed at runtime.
Add start9/0.4/render-smoke.mjs: a jsdom render smoke check that (1) runs the
shipped Babel over the app's inline JSX and asserts a classic, non-module,
parseable script (the v79 ESM-import regression), and (2) mounts the app in
jsdom and asserts the login UI renders (the v78 blank-screen class). Wired into
the default `make` goal so every package build is gated on the frontend
actually rendering — closing the "verified live via curl only" gap. jsdom is a
build-time devDependency, not shipped in the image.
The email-activity panel surfaced every captured message, including cold/
unknown-sender email with no investor association. Gate query_email_activity
on EXISTS(email_investor_links) so the panel shows only email tied to a known
investor/contact. Capture is unchanged — unmatched email is still stored
(metadata-only) and will appear automatically if its sender is later added as
an investor; this is a read-side filter only.
Graveyard investors are unaffected (their email has a link), so they remain
visible/searchable as an audit surface, hidden only from the filter picker.
The Communications tab is now an admin-only search over captured Gmail
(email_* tables), part of consolidating on the fundraising grid + email
capture as the canonical system of record.
- New GET /api/email/activity (admin-enforced server-side): filter by
investor / mailbox / direction with free-text search over subject,
snippet, and sender. Query logic in db.query_email_activity.
- Soft-delete honored on the per-mailbox sighting (emails carry no
deleted_at; deletion lives on email_account_messages).
- Direction decided at the email level (outbound if the sender is one of
our mailboxes), mirroring digest_builder.
- Graveyard investors are hidden from the filter dropdown (CRM-wide
graveyard=0 convention) but their email stays visible in the list and
findable by free-text search — this is an audit surface.
- Communications page rewritten to render the panel; the classic manual
"Log Communication" form is retired (the grid context menu remains the
manual-log path). Nav item + page are admin-only.
- Tests: email_integration/test_email_activity_panel.py (filters,
per-sighting soft-delete, roll-ups, graveyard handling, route 401/403);
full suite 22/22. Frontend render verified via a jsdom mount smoke test
plus the pinned classic-runtime Babel transform.
Code-only, no schema migration (version migrations are no-ops).
The web UI rendered a blank screen for every user. Root cause: the page
loaded @babel/standalone from unpkg with no version pin, so the CDN silently
served Babel 8.0.0. Babel 8 defaults @babel/preset-react to the automatic JSX
runtime, which prepends `import {jsx} from "react/jsx-runtime"` to the compiled
output. An ESM import is illegal in this classic (non-module) inline <script>,
so the browser rejected the whole bundle and React never mounted — hence the
blank screen. The prior "verified live" checks were server-up/curl, which can't
catch a browser-render failure.
- Pin @babel/standalone@7.29.7 (its preset-react defaults to the classic
React.createElement runtime). Verified via headless render: app mounts, login
screen renders, no console error. Follow-up: vendor + SRI-pin the CDN libs so
a third party can't swap our front-end deps in production again.
- Close three server-side admin gaps surfaced by a permissions audit — endpoints
that were UI-hidden from members but not API-enforced: GET /api/users,
/api/email/status, /api/email/accounts now require_admin. Removed the now-dead
non-admin mailbox-row filter. 21/21 backend tests green; py_compile clean.
The fundraising grid + email capture is the canonical system of record. lp_profiles
was a superseded single-fund model with no reachable create/edit path, and the LP
Tracker page was already orphaned (no nav entry + a redirect bouncing it to the grid).
- Remove /api/lp-profiles* endpoints + handlers, the unused lp-breakdown report,
the contact-dossier LP section, the demo-seed LP block, and (frontend) the
LPTrackerPage component + its lp-tracker->fundraising-grid redirect.
- Dashboard "Total Committed" now sums fundraising_investors.total_invested
(graveyarded investors excluded) instead of the orphaned lp_profiles table, which
read ~$0. "Total Funded" dropped: the grid tracks commitments, not a funded amount,
and the frontend never rendered it.
- Leave the empty lp_profiles table/index, the contact-delete soft-delete cascade,
and the --reset-all-data clear in place (never-hard-delete).
- Tests: add test_dashboard_report.py; update test_soft_delete_reads.py. 21/21 green.
Sends a once-a-day internal email to all active admins summarizing each team
member's email activity per investor, plus a team-wide by-investor view
(inbound + outbound, deduped). Narratives are generated on the LOCAL Spark
model, never Claude — the digest is intentionally un-anonymized, so substance
stays on Ten31 infra. This is an internal ops email, exempt from the
'agents draft, humans send' rule (which governs outward LP contact).
- backend/digest_builder.py: per-user + per-investor activity queries
(soft-delete filtered), per-user Spark narrative with a deterministic
fallback, two-section plain-text body, and the DB-backed policy resolver.
- backend/email_integration/digest_scheduler.py: always-on daily thread that
re-reads the policy each cycle and sends once/day; window cursor in
app_settings so a missed day rolls forward.
- server.py: POST /api/admin/digest/send-now and GET/PATCH
/api/admin/digest/policy; scheduler wired into main().
- Control lives in Settings -> Admin (enable toggle + send-time dropdown),
not StartOS actions; env vars only seed the first-boot default.
- Tests: backend/test_digest_builder.py.
The box's existing service-account domain-wide-delegation grant already includes
gmail.compose, which authorizes users.messages.send — verified 2026-06-15 by a
token-mint probe and a live messages.send to grant. So CRM-originated mail can
send through the account that already powers email capture: no SMTP account, no
app password, no admin change.
- backend/email_integration/gmail_send.py: send_via_gmail() impersonates a
domain user and POSTs users.messages.send (reuses credentials.py + the compose
scope; mirrors compose.py's REST pattern).
- backend/digest_mailer.py: send_digest() prefers Gmail DWD when enabled, falls
back to smtp_send otherwise. Sender = CRM_DIGEST_SENDER else first active admin.
- server.py: the admin test endpoint now routes through digest_mailer (so the
Settings button sends via DWD on the box with zero SMTP config). Recipient
restriction to the admin set and no-leak error handling preserved.
- test_gmail_send.py: build/send + transport routing (provider + urlopen faked).
19/19 backend green; s9pk typechecks.
SMTP (v75) stays as the fallback transport. Send-path decision + scope finding
recorded in ROADMAP.md and AGENTS.md.
Surface the digest test-send endpoint as a clickable admin control so it can be
exercised on the box without curl. Calls POST /api/admin/digest/test-email and
toasts the result (or a 'configure SMTP first' hint). JSX parse-checked.
Groundwork for the daily activity digest: give the CRM an outbound mail path.
Today nothing leaves the box (Gmail capture + drafts only), so this adds a
dedicated, per-package SMTP account independent of any StartOS system-wide SMTP.
- configureDigestSmtp Start9 action: writes host/port/from/username/password/
security to /data/secrets/smtp/* (password piped over stdin, never argv/env;
per-field files, owner-only) — mirrors the setAnthropicApiKey pattern.
- docker_entrypoint.sh reads those at boot and exports SMTP_* (operator env wins).
- backend/smtp_send.py: stdlib smtplib wrapper reading SMTP_* (one code path for
dev .env and the box); starttls/tls/none modes.
- POST /api/admin/digest/test-email (admin-only): proves the pipe. Recipients are
restricted to the active-admin set — an arbitrary `to` is rejected, so the
endpoint is not an open relay; send failures are logged, not echoed (an SMTP
auth error can carry the credential).
- Tests: test_smtp_send.py (sender), test_smtp_endpoint.py (gating + relay
restriction + no-leak). 18/18 backend green; s9pk typechecks.
Analysis/summarization for the digest body (Phase B) will run on Spark, never
Claude — the digest is deliberately un-anonymized. Decisions + Phase B plan in
ROADMAP.md.
Fixes from the 2026-06-12 full-eval (P0 + two P1s); code-only, no schema
change. Without these the "private CRM" premise was breachable on the LAN:
- P0: the /assets/ route joined the request path onto FRONTEND_DIR without
normalizing '..' (get_path/urlparse pass it through), so an unauthenticated
GET /assets/../../data/crm.db read any file the process could — the LP DB,
the JWT signing secret (-> admin-token forgery), the Gmail key. Add a realpath
containment check that 404s anything resolving outside FRONTEND_ROOT.
- P1: the LP-outreach drafter built its redaction Boundary with no ner_fn, so
unknown people/firms in raw email bodies reached Claude in the clear. Pass the
local-Qwen NER backstop (ner_fn=_ner_local), matching architect_grounding;
fails closed via the existing scrub_unavailable path if the local model is down.
- P1: get-by-id handlers leaked soft-deleted records by direct ID. Add
deleted_at IS NULL to every get-by-id path — contacts, organizations,
opportunities, lp_profiles — and to the nested related-data sub-selects in
the contact/opportunity detail payloads, matching the list-handler convention.
Bumps the package to v0.1.0:74 (utils.ts + versions/v0.1.0.74.ts + graph).
Full report in EVALUATION.md; remaining P2/P3 triaged in AGENTS.md Current state.
Swap the dead "scarcity as the connecting idea" / bitcoin-as-settlement
spine for the v2.0 reserve-asset spine (bitcoin = apex non-debasable
reserve asset; debasement = forcing function; AI = abundance engine;
throughline is an asset-value/capital-flow claim, not settlement; three
seams Energy<->Compute, Debasement<->Bitcoin, AI<->Data-Ownership)
everywhere it was still encoded in live code, the seed, and the docs.
- architect_agent.py / outreach_agent.py: both system prompts carried
"scarcity as the connecting idea" and shipped settlement framing into
every generated draft; rewritten to the reserve-asset spine.
- thesis_seed.py: THROUGHLINE, PILLAR_1, the AI/energy-operator segment
angle, and THESIS_V2 corrected and voice-cleaned (no em dash / "X, not
Y" / "bet"). PILLAR_2/3 (real revenue, founder access) kept.
- ensure_thesis_v2_promoted / revert_thesis_v2_promotion: make the v2.0
spine the working APPROVED spine and re-ground/clean the core nodes,
deployment-state-invariant (structural targeting, not body text) and
fully reversible (captures prior body/title/status/deleted_at). NODE
level only: never sets a thesis_version canonical (guardrail #4); no
hard deletes (guardrail #3). Wired into init_db after the v2 candidate
stage.
- docs/thesis-handoff.md replaced wholesale with the complete v2.0 doc;
Ten31_Agentic_Build_Plan.md + PHASE_1.md throughline glosses updated.
The v2.0 spine remains an unratified draft from the signal-engine
workstream: canonical freeze stays the partners' dual sign-off, and
Appendix-A conviction/exposure figures stay Grant's working read.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Incorporates the signal-engine workstream's v2.0 thesis correction: the spine is bitcoin
as the apex NON-DEBASABLE RESERVE ASSET (debasement = forcing function, AI = abundance
engine), NOT "infrastructure settles on bitcoin" (the settlement/payments claim — Strike's
payments thesis died in backtest). thesis_seed.ensure_thesis_v2_candidate stages the
v2.0 root/forcing-function, throughline, the verifiable-vs-contrarian decomposition, and
the 3 seams (Energy↔Compute, Debasement↔Bitcoin, AI↔Data-Ownership) as CANDIDATE nodes
under the core line (idempotent sentinel; provenance + "unratified, exposure unconfirmed"
on the section). Nothing canonical (guardrail #4). docs/thesis-handoff.md gets a
SUPERSEDED-spine banner pointing to v2.0.
NOT done (gated on partner ratification): the live THROUGHLINE/PILLAR_1 constants and
architect_agent.py's system prompt ("scarcity as the connecting idea") still encode the
old spine — until ratified+updated, Vary/Revise/outreach regenerate the old framing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
(1) Voice: _voice_examples now picks the sender's prior sent emails OF THE SAME PURPOSE
(PURPOSE_PATTERNS keyword cues per outreach type), larger sample (8) weighted by purpose
then recency — not just recent. meta carries on_topic for transparency.
(2) Tier-B sending (gmail.compose now authorized in Workspace DWD). New
email_integration/compose.py create_outreach_draft: mints a compose-scoped DWD token for
the sender (credentials._mint/access_token_for parameterized by scope; GMAIL_COMPOSE_SCOPE),
builds an RFC822 message, and POSTs gmail.drafts.create into the SENDER's mailbox — as an
in-thread reply (threadId + In-Reply-To/References, recipient = matched LP address) when
there's an active thread, else a fresh email. NEVER sends — the human sends from Gmail
(guardrails #4, #6). Route POST /api/outreach/gmail-draft; UI "Create Gmail draft" button +
"Open Gmail Drafts" link. Tests: test_compose.py (parse/reply-target/RFC822+threading).
Message construction unit-verified; the live drafts.create runs on the box.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Voice upgrade. draft_outreach now learns the SENDER's voice: the codified rules PLUS a
few-shot of that user's own recent sent emails (_voice_examples; from_email = the
sender, de-identified in the same scrub batch as the recipient context, reference-only).
The response returns which of the sender's emails were used (subject + date + recipient),
shown in the UI as "Voice based on: …" — transparency to avoid the black-box problem.
Falls back to rules-only with a clear note when the user has no captured sent email.
Context restructured: _context groups the investor's email by thread and labels the most
recent thread as the "Active conversation (what you are replying to)" with earlier emails
as background, so replies stay on-topic instead of dredging old threads.
Sender email resolved in handle_outreach_draft (users table by user_id). Test extended
(active/background split, voice examples + meta, no-sender fallback). Fixed a UI bug the
preview caught: the manual Draft button was onClick={draft}, which passed the click event
as the investor arg after draft() gained params -> circular-JSON error; now onClick={()=>draft()}.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The Outreach page now opens with a "Needs attention" list. A deterministic scan
(outreach_agent.follow_up_radar) surfaces investors per the email history: tier 0 "you
owe a reply" (their email is the most recent, unanswered, >=3d), tier 1 flagged + quiet,
tier 2 warm lead gone quiet (no contact in >=45d). Most urgent first; every reason is
verifiable from the data (no LLM in the surfacing — the deliberate fix for the trust
problem that sank objection-grounding). Excludes graveyard; needs email history. One
click sets the investor + suggested type (follow-up/nurture) and runs the existing
outreach drafter. Route GET /api/outreach/radar. Test mcp/test_outreach.py extended
(owe-reply/warm-quiet/recent/graveyard/order). Verified live in preview.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
First proactive-messaging build. New "Outreach" page (all authenticated users): pick an
investor + type (intro / follow-up / fund update / meeting follow-up / nurture) + optional
guidance; the agent drafts a tailored LP email in Ten31's voice, grounded in the thesis +
that investor's CRM notes and matched email history. The draft is editable + copyable;
nothing is sent (draft-only — guardrails #4, #6).
Sovereignty: the thesis is Ten31's own non-sensitive messaging (to Claude as-is); the LP
context is scrubbed through the redaction boundary before Claude, drafted with placeholders,
and re-hydrated locally — the LP list never reaches the API. Fails closed (scrub_unavailable /
claude_not_configured / rehydrate_failed quarantines a hallucinated-token draft).
Backend: mcp/outreach_agent.py (context assembly + scrub + Claude + rehydrate, reusing
architect_agent's client/thesis/voice + the Boundary); routes GET /api/outreach/investors,
POST /api/outreach/draft; logged. Test mcp/test_outreach.py (context assembly). Verified in
preview: page/selector/types/guidance render, fail-closed at the key-less Claude step (scrub
ran locally first), success rendering verified with a mocked ok draft.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The summarize-historical-email grounding produced generic, boilerplate objections
with no quotes and no source traceability (the minimize step abstracts away the
actual email text; the newest-N corpus carries little real objection signal, so the
model pattern-completes). Pulled the page (ObjectionsPage component + nav + dispatch).
The redaction boundary is kept (reusable for proactive outreach); the dormant
/api/architect/ground route is left in place but has no UI trigger. Pivoting to
proactive outreach / messaging.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New admin "LP Objections" page (frontend ObjectionsPage + nav). Pick a segment (or
All LPs) and Run grounding: the Architect mines matched LP emails + notes on the local
model, scrubs every identifier through the redaction boundary, and asks Claude for the
recurring objections + honest rebuttals (substantiated/hand-wavy flagged). Renders the
de-identified draft + an "N identifiers protected" badge; fail-closed statuses
(local_model_unavailable / scrub_unavailable / claude_not_configured / rehydrate_failed)
show a clear message. Uses the existing /api/architect/ground route. Verified in preview:
page + segment selector + Run; the local minimize/scrub legs actually ran against real
Spark on synthetic input and fail-closed correctly at the (key-less) Claude step;
success rendering verified with a mocked ok response.
NOT yet deployed — start-cli RPC to the box hit a transient transport error post a
StartOS hiccup (curl works, start-cli doesn't); CRM healthy at v0.1.0:65 meanwhile.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
/api/email/accounts now returns captured + matched per account (from the per-mailbox
sighting table email_account_messages joined to emails; emails dedupe globally so an
email seen by two mailboxes counts for each). Each mailbox card on the Email Capture
page shows "<N> captured · <M> matched" so per-user coverage is visible, not just the
aggregate. Verified in preview with two seeded mailboxes.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
When a sent/received email is matched to an investor, a local-model agent drafts a
one-line dated note and queues it as a PENDING proposal (it never writes the grid
itself). On the Email Capture page a partner sees "Proposed grid notes", can edit the
text, and Approve (appends to that investor's grid notes cell, newest at bottom,
stamped with the approver) or Dismiss. Going-forward only: a cutoff (app_settings
email_activity_since, set on first run) means email dated before the feature was
enabled is never summarized, so the historical backfill makes no noise. Sovereign:
summaries run entirely on the local model (no redaction needed). Gmail sync interval
tightened 180 -> 15 min so outgoing email surfaces quickly.
Backend: migration 0002 (email_activity_proposals); propose_email_activity_notes()
runs via a new scheduler post_sync hook; list/decide functions + routes
GET /api/activity/proposals, POST .../{id}/approve|dismiss. Grid append stamps the
approving user (fundraising_state.updated_by has a FK to users). Test
test_email_activity.py (propose cutoff/idempotency, approve appends + edited note,
dismiss, already-decided guard) under FK enforcement.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
/api/system/status now returns a best-effort storage block: database file size
(crm.db + WAL + SHM), the email_attachments dir, the backups dir, and disk
total/used/free via shutil.disk_usage(DATA_DIR). System Status renders a Storage
section with human-readable sizes so growth can be watched over time.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
insert_email's recipients loop did `for a in parsed.get(kind, [])`, but the parser sets
reply_to=None when there is no Reply-To header, so .get returns None (key present) and the
loop raised 'NoneType' object is not iterable — aborting the entire Gmail backfill on the
first such email (i.e. almost immediately). Fixed with `or []`. Regression test
test_insert_email.py (reply_to=None, all-None recipients, happy path).
Because the scheduler intentionally skips error-status accounts (no retry storms), an
errored mailbox would never resume on its own. "Sync now" now clears error status first,
so it is an explicit retry; backfill resumes from its saved cursor and dedups by
Message-ID, so nothing is re-captured.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The first Gmail backfill leaves the account at "pending · never synced" until it
fully completes (the sync_runs row only finalizes at the end), so there was no
feedback. /api/email/status now also returns captured_emails (total, which climbs
page-by-page during backfill), the latest sync run, and a backfilling flag. The
panel shows a "Backfilling… N captured so far" banner + an Emails Captured count
and auto-refreshes every 5s while a backfill is in progress. Verified live in
preview with seeded data (count auto-climbed 37 -> 50 without manual refresh).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds a "Test with a single mailbox first" input (pre-filled with the admin's own
address) + Enroll this mailbox button calling the enroll-one endpoint, so capture
can be tried on one mailbox before enrolling the whole domain. runAction now sends
an optional JSON body. Enroll-all stays.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds an admin-only "Email Capture" page so Gmail capture can be turned on and
monitored from the UI instead of an API call: shows whether the integration is
enabled, how many mailboxes are enrolled, how many emails are matched to investors,
and last sync; with Enroll Ten31 mailboxes / Sync now / Re-match buttons and a hint
that domain-wide delegation must be authorized in Google Workspace first. Disabled
state renders cleanly (no scary error) when the integration is off. Bundles the
email-into-grounding corpus wiring (bf829b7).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Saves the 2026-06-05 Architect positioning pass as competing CANDIDATE options
under the core line's positioning variant group, beside Option A/B: Convergence
(47/60), Access (40), Asymmetry (36), Scarcity/chokepoints (35), Freedom-tech (28),
each with its red-team weakness inline. One-time, additive, non-canonical
(guardrail #4); idempotent via an interaction_log sentinel so a partner-deleted
option is never resurrected. ensure_positioning_framings runs after the v5 seed.
Test: test_positioning_framings.py (count/candidacy/idempotency/no-resurrection/log).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The currency-anchored amount regexes treated a single-letter magnitude
suffix (k/m/b) as optional but unbounded, so "$5,000,000 but" scrubbed to
"[AMOUNT_1]ut" — the 'b' of "but" was consumed as a 'billion' suffix. Add a
word boundary after _MAG on the three currency-anchored _AMOUNT_RES patterns
(range, symbol, ISO-code); the worded-amount pattern is unaffected. Money
still tokenizes in every case ($5m/$5b/$3-5M/USD 5,000,000); only the OUTBOUND
to-Claude text stops losing the leading letter of the following word.
Round-trips were already lossless.
Regression-locked by a round-5 section in test_scrub_leak.py; full redaction
suite (scrub_leak + reidentification + grounding_boundary) green. Packaged as
StartOS v0.1.0:57. Reported by the Spark gateway dev; gateway re-vendored
scrub.py verbatim for parity (same golden-file leak test gates both sides).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Addresses Grant's feedback that the Workshop was confusing and underbuilt (no delete,
no approve, redundant generate-vs-feedback panels, and a stray "0" on segment lines).
Backend (architect_tools.py + server.py routes/handlers):
- retire_node: soft-delete a node + its subtree (reversible). DELETE /api/thesis/nodes/{id}.
- choose_variant: 'Use this' — keep this option, soft-delete the others in its group,
mark it approved. POST /api/thesis/nodes/{id}/choose.
- upsert_thesis_node gains actor_type so a manual human edit is recorded as 'human'.
PUT /api/thesis/nodes/{id} edits a part's text directly.
- handle_approve_line: one-click 'approve as current' — records this admin's approval on
the line's in-review version (creating + submitting one from the live tree if none),
promoting to canonical at the required distinct-approval count. POST /api/thesis/lines/{key}/approve.
Frontend (ThesisWorkshop redesign):
- Merged the redundant "Generate options" + "Give feedback" panels into one "Ask the
Architect for options" box (revise was just generate-with-guidance).
- Per option: Use this / Edit (inline) / Delete. Per part: edit + delete via the same.
- "Approve as current" bar with dual-sign-off state + a "Current ✓" badge, and a one-line
"how it works" hint. Refreshes the tree after every action.
- Fixed the stray "0": `{line.is_core && <badge>}` rendered 0 for non-core lines (SQLite
integer 0); now `{!!line.is_core && ...}`.
Verified: backend test_thesis_actions.py (choose/edit/retire-subtree/dual-approval->canonical),
and a live in-browser smoke test (JSX compiles, Workshop renders, options show Use/Edit/Delete,
approve returns 1-of-2, no runtime errors).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Phase 1 Workstream D. Lets the Architect ground the thesis in REAL recurring LP
objections without any LP identity reaching the Claude API. Layered, defense-in-depth,
fail-closed by construction (docs/redaction-rehydration.md).
backend/redaction/:
- scrub.py: the leak-proof core. Drops Tier-1 (labelled/structured account/wire/SSN/
IBAN/SWIFT/passport, separator-tolerant); tokenizes known LP entities (dictionary from
the canonical layer, unicode-folded + hyphen-extended) and structured PII (emails,
scheme-less/social URLs, intl+ext phones, currency-cued amounts, ISO/worded/numeric/
quarter dates, addresses, bare long digit runs); pre-neutralizes injected [TYPE_N]
strings; single-pass rehydrate; metadata-only audit logging (the pseudonym map is the
de-anon key — local-only, never logged/sent). Hardened across THREE adversarial
leak-hunts (worded/coded amounts, intl phones, NFD/ligature/zero-width names, slash/
comma SSN, SWIFT, alpha-prefixed accounts, substance-preserving false-positive fixes).
- client.py: Boundary — one scrub/rehydrate contract, SCRUB_BACKEND=local (default) or
gateway (Spark Control /scrub + /rehydrate). Fails closed (db_path required; dictionary
build errors propagate; strict rehydrate returns tokenized-not-de-anon text).
- test_scrub_leak.py, test_reidentification.py: golden-file leak + re-identification
suites (synthetic only, guardrail #9), regression-locking every leak-hunt vector.
backend/mcp/architect_grounding.py: the flow — retrieve (local) -> minimize-first
(local Qwen) -> scrub (+ local-Qwen NER backstop for unknown names) -> Claude over the
de-identified register only -> re-hydrate locally -> human review. FAILS CLOSED if the
local model is unreachable or a hallucinated token appears. test_grounding_boundary.py
proves nothing sensitive reaches Claude and the three fail-closed paths.
server.py: POST /api/architect/ground (admin) wires retrieval -> ground_objections.
docker_entrypoint.sh: SCRUB_BACKEND (default local). docs/spark-control-scrub-endpoints.md:
the gateway handover spec (Option 1 — caller supplies the entity dictionary).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The fundraising grid's per-contact editor now has a LinkedIn URL field next to
name, email, title, and location. It threads through the grid contact object and
sanitize (which preserves contact-object fields), and _upsert_contact_from_fundraising
now reads and persists linkedin_url on both the update and insert paths — so a
LinkedIn entered in the grid lands on the linked contact record.
Test: test_grid_contact_link.py extended to assert LinkedIn entered in the grid
persists to the contact (idempotent). Frontend html.parser clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
backend/thesis_seed.py builds the starting "living messaging source of truth"
from docs/thesis-seed-v5.md: a core line (throughline; the open Option A/B banner
as a competing variant group; the three pillars; the proof; voice rules), one line
per LP segment carrying that segment's angle, and the five segment definitions.
ensure_thesis_seed(conn) runs from init_db, seeding ONLY when the Workshop is empty
(no thesis lines) — idempotent and non-destructive, so it bootstraps once and never
overwrites partner edits. Everything lands draft/candidate; nothing is made canonical
(that stays the partners' dual-approval action, guardrail #4). Content is Ten31's own
messaging, not LP data.
Test: backend/test_thesis_seed.py runs init_db and asserts the core line, 5 segment
lines, the 2-member Option A/B variant group, 3 pillars, segment_cuts, and segment
defs, plus re-seed-is-a-no-op (11/11).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Structural fix for the duplicate-people class of bug: instead of matching a grid
contact "pill" to a contacts row heuristically by name/email (which drifted and
caused the 1406 double-count), link them by id.
Backend:
- Migration 0004: fundraising_contacts.contact_id (additive, nullable, logical FK
to contacts(id)) + index. Paired down migration.
- sync_fundraising_relational now stores the id that _upsert_contact_from_fundraising
already returns, so every grid contact carries its contacts-table id.
- _backfill_grid_contact_ids: one-time, idempotent backfill on startup (re-runs the
grid sync once if any row lacks contact_id), so existing data links immediately.
- entity_resolution: grid pass prefers the explicit contact_id link (match_kind
'grid_link') over heuristic email / name+investor, guarded by a PRAGMA check so
older DBs without the column still work.
Frontend:
- Fundraising grid "+ Row" -> "+ Investor" (clear, single investor entry point).
- Contacts page: the "+ Add Contact" trigger is replaced by a pointer to the grid;
the page is now a read/search/edit view (ContactDetailPanel still edits all
fields). New people are added from the grid. No contact data is removed.
Tests: backend/ingest/test_entity_resolution.py extended (explicit-link case, 11/11)
and a new backend/test_grid_contact_link.py integration test (init_db applies 0004,
sync populates contact_id to the right contact, re-sync is idempotent). py_compile +
frontend html.parser clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Root cause: grid contacts (fundraising_contacts) are the SAME people as the
contacts table (the app syncs them by name/email), but resolution matched grid
rows by (name + investor-canon) where the two sides derive the investor key from
different tables that rarely line up — so nearly every grid contact minted a
duplicate person (715 + ~692 ≈ 1406), and the duplicate finder then flagged each
twin against its real self (~676 candidates).
Fix (entity_resolution.py):
- Grid pass matches a grid contact to its existing contacts-table person by
PROVABLE keys only (exact email, else exact name within the same investor) and
records membership; on a miss it MINTS NOTHING (the old else-branch mint was the
double-count source, and guessing by name across firms risks binding two
different same-named people).
- Targeted, audited cleanup soft-deletes leftover grid-only "twins" (person rows
with no 'contacts' link) and superseded pre-:48 'lp'/'organization' rows, guarded
so any row carrying enrichment/human data is never dropped (guardrail #3); the
tombstoned ids are logged to interaction_log (guardrail #5).
- _upsert_entity clears deleted_at on conflict so a re-emitted id is un-tombstoned
(no permanent burial); fuzzy-merge losers stay buried via _redirect.
entity_merge.py / server.py: the duplicate queue + pending count now filter to
candidates whose both sides are still live, so self-healed twins drop out.
Verified: offline reproduction test (backend/ingest/test_entity_resolution.py,
10/10) reproduces the 1406-style doubling and proves it collapses; no regression
on the synthetic dev set; two adversarial review passes. Known pre-existing
identity-key weaknesses (same name+firm+no email collision; shared role inbox
over-link) are unchanged by this fix and will be resolved structurally by the
contact_id link in the grid/contacts unification.
Run "Build search index" after upgrading to recompute the canonical layer.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Lets a non-technical operator install the Architect's Claude key from the
StartOS UI instead of the terminal: a masked text field whose value is written
to /data/secrets/anthropic-api-key (0600) on the box — the same file the
entrypoint already loads at boot. Secret is piped over stdin (never argv/env),
CR/LF stripped to match the entrypoint's read. allowedStatuses 'any'; a restart
is required (and stated in the action's warning + success message) since the
entrypoint reads the key only at startup.
Verified the Architect's data boundary first: the deployed Thesis Workshop
routes send only Ten31's own thesis text (thesis_lines/thesis_nodes) + the
partner-typed guidance to Claude — no contacts/lp_profiles/communications/grid.
(The MCP CRM-retrieval tools that DO return record substance are not wired into
the deployed Architect; the redaction boundary must land before any grounding
path uses them — Phase 1 Workstream D.)
tsc --noEmit clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Frontend: ThesisWorkshopPage / ThesisWorkshopNode / ThesisWorkshopOptions —
the collaborative iteration screen where partners generate a variable number
of competing thesis options (1, 2, 3, A1/A2/A3 ...) for any node, give
feedback, and regenerate. Reuses the shared api() helper; flexible option
count is the core UX constraint.
Backend Architect agent (architect_agent.py) + routes shipped in dd25bbc;
this completes the user-facing surface and bumps the StartOS package to
0.1.0:49 (anthropic dep already in the image, key loaded from
/data/secrets/anthropic-api-key — self-disabling until present).
Also lands thesis seed iterations v3 and v5 (voice/messaging corrections).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Per Grant's clarification of the real data model:
- Investor entities come from the fundraising grid, one per row, all labeled
"investor" (drops the confusing lp/organization split). Grid is source of truth.
- People come ONLY from the contacts table. The grid's contacts (fundraising_
contacts) are matched to a contact-person and recorded as member_of links to
their investor, instead of creating duplicate person entities. This fixes the
~doubled people count (people now ≈ contacts, not contacts + grid contacts).
- System Status cards: Investors / People (resolved) / Contacts in CRM / Grid
contacts, so resolved-vs-source is visible at a glance.
Verified on synthetic: people == contacts count (no double-count); multi-contact
investors preserved via member_of.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The image COPY'd backend/server.py + a few subdirs but missed core_migrations.py,
backend/migrations/, and the Phase-1 modules (thesis_review/entity_merge/
entity_jobs). On the box the migrations never ran (tables absent) and those
endpoints 503'd ("Jobs unavailable"). Now COPY backend wholesale (.dockerignore
keeps __pycache__/data out). Bump to 0.1.0:46.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- frontend: System Status page extended with one-click index actions
(update/rebuild/find-duplicates, with live job status) and a human-in-the-loop
duplicate-review queue (approve=merge / reject=keep-separate per candidate).
- StartOS version 0.1.0:45 (image-only; schema via the in-app migration runner).
Backend + new routes verified end-to-end via the running HTTP server.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- backend/ingest/sync_scheduler.py: periodic incremental-sync loop (every
CRM_INGEST_SYNC_INTERVAL_MIN min); resilient, --once for testing.
- start9/0.4: "Refresh search index" action (incremental sync.py); entrypoint
launches the scheduler as a background process when Spark/Qdrant are set;
CRM_INGEST_SYNC_INTERVAL_MIN env; pre-release note on fastembed/mcp pins.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Fuzzy tier (backend/ingest/fuzzy_resolve.py + llm.py): local Qwen adjudicates
the deterministic resolver's flagged name-variant candidates; merges are
durable via entity_merges (deterministic re-runs respect them), losers
soft-deleted, logged. Idempotent.
- Incremental sync (backend/ingest/sync.py): re-embeds only rows changed since a
watermark (ingest_sync_state); first run / --recreate = full. Tested full→0→1.
- Start9 packaging (start9/0.4): Dockerfile bundles ingest+mcp + fastembed/mcp;
"Build search index" action runs the init in a subcontainer; MCP shipped as a
manual stdio server (not a daemon); version 0.1.0:44. INGEST_PACKAGING.md.
- backfill.py: factored embed_and_upsert() shared with sync.
Verified end-to-end on synthetic data + live Sparks/Qwen/Qdrant.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Keysat