Agent Data Hub v0.1.16

Agent Data Hub v0.1.16

Agent Data Hub v0.1.16 is a small operational-stability polish release after
the v0.1.15 public-demo doctor fix.

What Changed

• scripts/agent_preflight.sh now prints ADH-root-aware diagnosis and start
  commands when the durable Hub appears offline.
• Docker timeout, missing-container, stopped-container, and pending-migration
  hints now point to the actual ADH checkout scripts instead of relative
  scripts/... paths that can be wrong from another project.
• scripts/db_doctor.sh and scripts/db_recover.sh now show root-aware
  recovery/start commands in operator messages.
• Regression tests keep these offline and recovery hints usable from
  non-ADH working directories.

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no automatic recovery
• no change to backup or recovery behavior

Release Position

v0.1.16 does not change what Agent Data Hub stores or writes. It makes the
existing local operations path less confusing when agents call ADH wrappers
from another repository and the Hub is unavailable or not ready.

Agent Data Hub v0.1.15

Agent Data Hub v0.1.15

Agent Data Hub v0.1.15 is a small public-demo reliability release after the
v0.1.14 seed-boundary polish.

What Changed

• scripts/db_doctor.sh --public-demo now diagnoses the same isolated demo
  database used by scripts/first_run_demo.sh.
• The public quickstart and first-run walkthrough now point demo users to that
  demo-aware doctor command instead of the configured operator doctor path.
• Regression tests keep the public demo troubleshooting path separate from the
  maintainer-local operator database.

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no hosted service
• no automatic recovery
• no change to the public first-run command

Release Position

v0.1.15 keeps the same technical-preview scope as v0.1.14. It does not add a
new feature; it closes a first-run clarity gap found by a clean-clone test:
public demo troubleshooting now targets the public demo database explicitly.

Agent Data Hub v0.1.14

Agent Data Hub v0.1.14

Agent Data Hub v0.1.14 is a small public-preview hygiene release after the
v0.1.13 first-run compression.

What Changed

• seed/README.md now explains the difference between the neutral public demo
  seed and the maintainer-local operator seeds.
• The README Startup Paths section links to that seed boundary so outside
  developers can understand why local maintainer paths exist in the repository.
• A regression test keeps the public demo seed path separate from maintainer
  local ops seed files.

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no hosted service
• no automatic promotion
• no change to the public first-run command

Release Position

v0.1.14 keeps the same technical-preview scope as v0.1.13. It does not add a
feature; it makes the public repository easier to read by naming which seed data
is neutral public demo seed material and which seed data belongs to the
maintainer's local operator workflow.

Agent Data Hub v0.1.13

Agent Data Hub v0.1.13

Agent Data Hub v0.1.13 is a first-run and local-operations hardening release
for the local-first technical preview.

It keeps the product claim narrow: reviewed context for humans and agents, not
a hosted service, not an autonomous memory grabber, and not a replacement for
repo-local working rules.

What Changed

Public First-Run Compression

• The README quickstart now points new users through one short path instead of
  presenting several equal-looking startup routes.
• The public getting-started guide now separates the fast path, expected first
  checks, troubleshooting, mobile preview, and manual path.
• A new first-run demo session documents the exact happy path: start the demo,
  open the demo project, review one draft, prepare one agent handoff, and check
  the boundaries.

Local Ops Reliability

• agent-hub doctor diagnoses the local Docker/Postgres Hub runtime without
  writing memory.
• scripts/db_doctor.sh checks Docker, Compose, the configured container and
  volume, Postgres readiness, recent logs, agent-hub status, and
  agent-hub check.
• scripts/db_recover.sh --apply provides a guarded recovery path for known
  stale Postgres lock-file failures.
• Recovery creates a Docker-volume snapshot first, recreates only the
  container, removes only empty or NUL-only postmaster.pid, and never removes
  volumes or alters Hub memory rows.

Backup Health Friction Reduced

• Regular preflight still requires a fresh local backup with a valid checksum.
• Configured remote backup parity is still checked and reported.
• Remote parity becomes a hard blocker only with
  AGENT_HUB_REQUIRE_REMOTE_BACKUP=1 or
  scripts/db_backup_health.sh --require-remote.

Recommended First Run

From a fresh folder:

git clone https://github.com/AlexanderSmyslowski/central-agent-data-hub.git
cd central-agent-data-hub
scripts/first_run_demo.sh

Then follow:

docs/public/first-run-demo-session.md

Verification

This release candidate was checked with:

git diff --check
bash -n scripts/*.sh
.venv/bin/python -m compileall agent_hub
.venv/bin/python -m pytest -q
agent-hub doctor
.venv/bin/python -m agent_hub.cli doctor
scripts/db_backup_health.sh --require --require-remote
scripts/db_recover.sh
scripts/smoke_public_demo.sh
agent-hub status
agent-hub check

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no hosted service
• no automatic recovery without explicit --apply
• no autonomous agent execution
• no authentication or role system

Release Position

v0.1.13 is suitable for evaluating whether a new developer can understand the
core loop quickly: start a neutral local demo, inspect reviewed memory, review a
draft explicitly, prepare an agent handoff, and diagnose local runtime problems
without guessing.

It remains a technical preview. The release improves first-run clarity and
local reliability; it does not broaden Agent Data Hub into a hosted product or
an automatic memory system.

Agent Data Hub v0.1.12

Agent Data Hub v0.1.12

Agent Data Hub v0.1.12 is a Hub View app-surface release for the local-first
technical preview.

It continues the same direction as v0.1.11: make reviewed project memory more
visible, easier to inspect, and easier to hand to a chatbot or local agent
without pretending that Agent Data Hub is a hosted product or an autonomous
agent runner.

What Changed

Hub View Feels More Like A Local App

• Hub View now has clearer app navigation, active-project state, project-area
  links, and mobile section navigation.
• Project pages show a stronger work center: current state, next actions,
  review queue, quality signals, memory access, and agent handoff.
• The public demo uses more direct first-run paths so a new user can open the
  app and understand what to inspect next.
• The UI remains a local surface over the Hub database. It is not the operative
  source of truth.

Memory Library And Detail Pages

• Reviewed memory can now be browsed through a dedicated memory library.
• Facts, decisions, risks, reports, and open questions have clearer labels,
  counts, and detail pages.
• Detail pages show useful actions: return to the library, filter by type,
  continue from that memory, and prepare an agent handoff from the item.
• Search feedback is more visible: matches appear while typing, empty sections
  are hidden, and reset actions are clearer.

Review Flow Visibility

• Review Inbox cards explain what is being remembered, where it came from, and
  what could go wrong if the item is wrong.
• Accept and reject outcomes are easier to understand after the action.
• A review history page shows recent human decisions as a read-only audit view.
• Reviewer identity is still attribution, not authentication.

Agent Handoff Is Easier To Start

• The agent handoff page now starts with a short Start here panel.
• The panel points to three practical paths: choose an agent path, check that the
  handoff worked, or copy text for a chatbot.
• Codex, Claude Code, Hermes/custom agents, MCP-compatible agents, chatbots, and
  terminal fallback remain separate paths with honest checks.
• Hub View shows the handoff and counts. It does not run an agent and cannot
  prove that an unconnected agent used the context.

Mobile And Language Polish

• Mobile-first navigation is more visible across project, memory, review, and
  agent-handoff surfaces.
• The German Hub View surface now covers more labels, controls, empty states,
  review text, and handoff guidance.
• Some domain terms remain in English where they are the common tool term, but
  the surrounding user guidance is clearer.

Code Structure

• Hub View templates and static assets are split into smaller modules.
• CSS and JavaScript are organized by surface instead of being concentrated in a
  single page file.
• The static asset manifest is guarded so frontend assets cannot silently drift
  out of the served app.

Recommended First Run

From a fresh folder:

git clone https://github.com/AlexanderSmyslowski/central-agent-data-hub.git
cd central-agent-data-hub
scripts/first_run_demo.sh

Then open Hub View and check:

• the project work state
• the memory library
• one memory detail page
• the Review Inbox and review history
• the agent handoff page, starting with Start here

For a phone preview on the same trusted Wi-Fi:

scripts/first_run_demo.sh --mobile

Verification

This release candidate was checked with:

git diff --check
bash -n scripts/*.sh
.venv/bin/python -m compileall -q agent_hub
node --check templates/hub_view/static/*.js
.venv/bin/python -m pytest -q
scripts/smoke_public_demo.sh
.venv/bin/python -m agent_hub.cli status
.venv/bin/python -m agent_hub.cli check

Visible Hub View changes were also checked in a real browser at desktop and
mobile sizes during the app-surface loops.

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no hosted service
• no autonomous agent execution
• no authentication or role system

Release Position

v0.1.12 is suitable for evaluating whether Hub View now works as a small local
app surface for reviewed project memory: something a person can open, inspect,
search, review, and use to prepare agent context.

It remains a technical preview. The release improves visibility, navigation,
review legibility, memory inspection, and agent handoff clarity; it does not
turn Agent Data Hub into a hosted product or replace the reviewed-memory model.

Agent Data Hub v0.1.11

Agent Data Hub v0.1.11

Agent Data Hub v0.1.11 is a small Hub View work-surface release for the
local-first technical preview.

It focuses on a simple first-run question: after the demo opens, can a person
see what they can do next and understand the current state of the reviewed
project memory?

What Changed

Clearer Project Actions

• Hub View now opens with a What can I do here? section.
• The section gives five concrete paths: use ADH with an agent, find reviewed
  memory, read latest status, review suggested changes, and check memory
  quality.
• The German UI mirrors the same idea with Was kann ich hier tun? and
  action-oriented wording.
• The page search recognizes common German UI terms such as Risiko,
  Entscheidung, Bericht, and Arbeitsstand when filtering reviewed memory.
• The actions point at existing read/review surfaces. They do not create new
  write paths or background automation.

Current Work State

• Project pages now show a compact Current work state block.
• The block summarizes latest status, risks and open questions, review queue,
  and quality snapshot.
• In German, the same section is shown as Aktueller Arbeitsstand.
• The goal is to make Hub View useful as a local overview, not just a static
  demo display.

Same Boundaries

• Hub View still reads from the local Hub database.
• Reviewed memory remains the durable source for facts, decisions, risks,
  questions, reports, and relations.
• Suggested memory changes remain unreviewed until a human accepts or rejects
  them.
• Agent handoff remains visible and explicit; the UI does not claim that an
  unconnected agent has used the context.

Recommended First Run

From a fresh folder:

git clone https://github.com/AlexanderSmyslowski/central-agent-data-hub.git
cd central-agent-data-hub
scripts/first_run_demo.sh

Then open Hub View and look for:

• What can I do here?
• Current work state
• Use ADH with an agent
• Find reviewed memory

For a phone preview on the same trusted Wi-Fi:

scripts/first_run_demo.sh --mobile

Verification

This release candidate was checked with:

git diff --check
bash -n scripts/*.sh
.venv/bin/python -m compileall agent_hub
.venv/bin/python -m pytest -q
scripts/smoke_public_demo.sh
.venv/bin/python -m agent_hub.cli status
.venv/bin/python -m agent_hub.cli check

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no hosted service
• no autonomous agent execution
• no authentication or role system

Release Position

v0.1.11 is suitable for evaluating whether Hub View now feels like a useful
local work surface for reviewed project memory, rather than only a demo page.

It remains a technical preview. The release improves orientation, visible next
actions, and current-state readability; it does not make Agent Data Hub a hosted
product or replace the reviewed-memory model.

Agent Data Hub v0.1.10

Agent Data Hub v0.1.10

Agent Data Hub v0.1.10 is a small Hub View language and legibility release for
the local-first technical preview.

It adds a simple bilingual UI layer for Hub View so first-run testers can switch
between English and German without changing stored memory, schema, or agent
handoff behavior.

What Changed

Hub View Language Switch

• Hub View now has a visible language switcher for English and German.
• The selected language is carried through project pages, the Review Inbox, and
  the agent handoff view with ?lang=de or the English default.
• The page <html lang> attribute follows the selected language.
• The switcher uses larger touch targets on mobile so language changes do not
  depend on precise tapping.

German UI Chrome

• Major Hub View labels, navigation, empty states, search controls, Review Inbox
  controls, and agent-handoff instructions are available in German.
• German UI labels use human terms such as Projektgedächtnis and
  Prüfungseingang where those are clearer than untranslated internal labels.
• Built-in public demo descriptions are localized, while stored memory content
  stays in its original language.
• Browser-side feedback such as the memory-search result line and copy-button
  state follows the selected language.
• The first-run project page keeps the human inspection flow visible in both
  languages: reviewed memory, risks and questions, latest status, search, and
  agent handoff.

Memory Content Stays Original

• Stored facts, decisions, risks, reports, relations, context packs, and audit
  messages are not translated by Hub View.
• The language layer is UI chrome only. It does not alter reviewed memory or
  create a second source of truth.

Recommended First Run

From a fresh folder:

git clone https://github.com/AlexanderSmyslowski/central-agent-data-hub.git
cd central-agent-data-hub
scripts/first_run_demo.sh

Then open Hub View. Use the language switcher in the header to move between
English and German.

For a phone preview on the same trusted Wi-Fi:

scripts/first_run_demo.sh --mobile

Verification

This release was checked with:

git diff --check
bash -n scripts/*.sh
.venv/bin/python -m compileall agent_hub
.venv/bin/python -m pytest -q
scripts/smoke_public_demo.sh
.venv/bin/python -m agent_hub.cli status
.venv/bin/python -m agent_hub.cli check

Boundaries

• no schema change
• no migration
• no new dependency
• no translation of stored memory
• no new Hub-memory write path
• no hosted service
• no authentication or role system

Release Position

v0.1.10 is suitable for evaluating whether Hub View feels more approachable for
German-speaking first-run users while keeping ADH's reviewed-memory model
unchanged.

It remains a technical preview. The release improves local UI language and
orientation; it does not make Agent Data Hub a localization platform or change
how agents receive context.

Agent Data Hub v0.1.9

Agent Data Hub v0.1.9

Agent Data Hub v0.1.9 is a small Hub View usefulness release for the
local-first technical preview.

It does not add a new memory model, hosted service, autonomous search engine,
or write path. The release makes the project page more useful for humans by
adding a simple way to find reviewed memory already visible on the page.

What Changed

Find Reviewed Memory

• Project pages now include Find reviewed memory near the top of the page.
• The filter works on the visible reviewed memory already loaded into Hub View:
  facts, decisions, risks, open questions, reports, latest status, and
  relations.
• Type names such as risk, decision, source, or report work because the
  filter checks item types as well as visible text.
• Matching items appear directly below the search field while typing.
• The result count updates while typing.
• Empty matches show a clear no-result message.

Honest Scope

• The filter does not search drafts.
• The filter does not query hidden data.
• The filter does not use embeddings, background indexing, or a hosted search
  service.
• It is a local browser filter over the reviewed memory shown on the page.

Better First Human Use

• Find reviewed memory is now a first-level project action.
• The page still keeps Use ADH with an agent visible, but the first human
  value is clearer: inspect and find the reviewed context before handing it to
  an agent.

Recommended First Run

From a fresh folder:

git clone https://github.com/AlexanderSmyslowski/central-agent-data-hub.git
cd central-agent-data-hub
scripts/first_run_demo.sh

Then open Hub View and use:

1. Find reviewed memory
2. Type into Search this page
3. Jump to Reviewed memory, Risks and questions, Latest status, or
   Relations
4. Use Use ADH with an agent when you want to hand that context to an agent

For a phone preview on the same trusted Wi-Fi:

scripts/first_run_demo.sh --mobile

Verification

This release was checked with:

git diff --check
bash -n scripts/*.sh
.venv/bin/python -m compileall agent_hub
.venv/bin/python -m pytest -q
scripts/smoke_public_demo.sh
.venv/bin/python -m agent_hub.cli status
.venv/bin/python -m agent_hub.cli check

The first-run observation was repeated on a phone preview. The tester could
find Find reviewed memory, use Search this page, see matching items
appear directly below the field, understand that the search only covers visible
reviewed memory, reset the view with Show all, and still find Use ADH
with an agent.

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no autonomous search or background indexer
• no hosted service
• no MCP write tools
• no authentication or role system

Release Position

v0.1.9 is suitable for evaluating whether Hub View feels more useful as a
human review surface: a user can inspect the project and quickly find reviewed
memory before using it with an agent.

It remains a technical preview. The release improves local inspection and
orientation; it does not turn Hub View into a full product dashboard or make
ADH a general search platform.

Agent Data Hub v0.1.8

Agent Data Hub v0.1.8

Agent Data Hub v0.1.8 is a small Hub View usability release for the local-first
technical preview.

It does not add a new memory model, hosted service, autonomous agent execution,
or write path. The release turns the visible agent handoff page into a clearer
setup flow: choose the agent, follow one setup card, then check whether the
handoff is visible.

What Changed

Connect Your Agent

• The agent-context page now starts the setup area with Connect your agent.
• The setup flow shows three explicit steps:
  • choose the agent
  • connect once where possible
  • check the handoff
• The page keeps the honest boundary: Hub View creates setup material and
  reviewed context for the agent; it never runs an agent by itself.

Agent Picker

• The page now has a Choose your agent picker with direct links for:
  • Chatbot
  • Codex
  • Claude Code
  • Hermes/custom agent
  • Other MCP-compatible agent
  • Terminal fallback
• Each picker card says what kind of setup is expected: copy/paste, one local
  click, one copied command, persistent rule, MCP config, or temporary terminal
  run.

Handoff Check

• A new Check the handoff block explains what the user should look for after
  setup.
• Chatbots should receive the copied text before the task.
• Local agents should show an ADH Context Loaded receipt or equivalent
  counts at task start.
• The terminal fallback remains clearly labelled as temporary and not the normal
  automation path.

Recommended First Run

From a fresh folder:

git clone https://github.com/AlexanderSmyslowski/central-agent-data-hub.git
cd central-agent-data-hub
scripts/first_run_demo.sh

Then open Hub View and use:

1. Use ADH with an agent
2. Prepare agent handoff
3. Connect your agent
4. Choose your agent
5. Check the handoff

For a phone preview on the same trusted Wi-Fi:

scripts/first_run_demo.sh --mobile

Verification

This release was checked with:

git diff --check
bash -n scripts/*.sh
.venv/bin/python -m compileall agent_hub
.venv/bin/python -m pytest -q
scripts/smoke_public_demo.sh
.venv/bin/python -m agent_hub.cli status
.venv/bin/python -m agent_hub.cli check

A short first-run observation confirmed that a tester can find Connect your
agent, use Choose your agent, understand Connection verification, and
explain Check the handoff.

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no autonomous agent execution
• no hosted service
• no MCP write tools
• no authentication or role system

Release Position

v0.1.8 is suitable for evaluating whether Agent Data Hub can guide a first-time
user from "I can see reviewed memory" to "I know how to hand this context to my
agent and how to check the handoff."

It remains a technical preview. The release improves orientation and
verification language; it does not guarantee that an unconfigured external agent
used the context.

Agent Data Hub v0.1.7

Agent Data Hub v0.1.7

Agent Data Hub v0.1.7 is a small Hub View app-readiness release for the
local-first technical preview.

It does not add a new memory model, a hosted service, or autonomous agent
execution. The release makes the visible agent handoff easier to find and easier
to trust: users can see how reviewed ADH context is handed to a chatbot or local
agent, and what setup state ADH can actually verify.

What Changed

Release Hygiene

• pyproject.toml now matches the public patch version.
• The release was checked with a short first-run observation before tagging.

Project Actions In Hub View

• Project pages now show a compact Project actions strip near the top.
• The first action is Hand context to an agent, which jumps directly to the
  visible context-pack flow.
• The Create context pack action is styled as the primary next step, with a
  visible task field instead of a quiet inline command.
• Other actions point to suggested memory changes, reviewed memory, and memory
  quality.
• The actions are links to existing Hub View surfaces. They do not create a new
  write path or change the memory model.

Clearer Agent Connection Paths

• The agent-context page now separates the main paths:
  • chatbot copy/paste
  • Codex one-click local setup
  • Claude Code setup command
  • Hermes/custom startup rule
  • generic MCP config shape
  • terminal fallback
• Codex is shown before Claude Code because Hub View can provide a guarded local
  setup action when the project folder is known.
• The terminal command is explicitly framed as a temporary fallback, not the
  intended daily workflow.
• Long commands and local paths wrap cleanly on desktop and mobile.

Connection Verification

• Hub View now shows a Connection verification block before the setup cards.
• The verification block is visually separated from the setup cards and uses
  short status language so the difference between checked and external setups is
  easier to scan.
• Codex setup has a real local status:
  • Codex setup verified
  • Codex setup not installed yet
  • Demo preview only
  • Cannot verify yet
• Hub View can verify Codex only by inspecting the repo-local AGENTS.md block
  in a registered project folder.
• Claude Code, Hermes/custom agents, generic MCP agents, and chatbots stay
  honestly labelled as manual or external checks. ADH does not claim it can
  inspect another agent's private configuration or prove that an unconnected
  agent read the context.

Recommended First Run

From a fresh folder:

git clone https://github.com/AlexanderSmyslowski/central-agent-data-hub.git
cd central-agent-data-hub
scripts/first_run_demo.sh

Then open Hub View and use:

1. Hand context to an agent
2. Create context pack
3. Connection verification

For a non-blocking check without starting Hub View:

scripts/first_run_demo.sh --no-hub-view

For a phone preview on the same trusted Wi-Fi:

scripts/first_run_demo.sh --mobile

Verification

This release was checked with:

git diff --check
bash -n scripts/*.sh
.venv/bin/python -m compileall agent_hub
.venv/bin/python -m pytest -q
scripts/smoke_public_demo.sh
.venv/bin/python -m agent_hub.cli status
.venv/bin/python -m agent_hub.cli check

The final short first-run observation confirmed that a tester could find
Hand context to an agent, use Create context pack, recognize
Connection verification, and explain the difference between chatbot
copy/paste, Codex local setup, Claude setup, Hermes/custom startup rules,
generic MCP config, and the terminal fallback.

Boundaries

• no schema change
• no migration
• no new dependency
• no new Hub-memory write path
• no autonomous agent execution
• no hosted service
• no MCP write tools
• no authentication or role system

Release Position

v0.1.7 is suitable for evaluating Agent Data Hub as a local technical preview
where the question is not only "what reviewed memory exists?" but also "can I
see how that context is handed to an agent, and what setup state is actually
verified?"

It remains a technical preview. This release improves Hub View guidance and
connection visibility; it does not turn ADH into a hosted app or guarantee that
an unconfigured external agent used the context.