Commit graph

3129 commits

Author SHA1 Message Date
Alessandro Frau
2b0b3a416d Add host workspace tree settings with container inheritance
Some checks failed
Build And Publish Docker Images / plan (push) Has been cancelled
Build And Publish Docker Images / build (push) Has been cancelled
Add Host workspace controls to Agent Settings > Workdir. Inherit the current
container workdir's four tree limits by default, or persist independent remote
limits when inheritance is disabled. Validate non-negative whole numbers and
preserve zero as unlimited without changing host access permissions.

Include effective remote_tree_config in connector hello and tree
acknowledgements so updated CLI and Launcher gateway sessions refresh their
maps without reconnecting. Older connectors remain compatible.

Document the settings and protocol behavior and add inheritance and validation
coverage. Verification: 109 focused Core tests passed; desktop/mobile UI,
Save/Cancel, persistence, and live gateway refresh checks passed. The full Core
suite was not run.
2026-11-17 15:28:13 +01:00
Alessandro Frau
5c94f04dea Add scheduled-task catch-up card to What's New
Introduce a second showcase card with the supplied Scheduled Tasks screenshot
and concise guidance on catching up recurring tasks after downtime. Make the
title explicitly refer to scheduled tasks so it cannot be confused with chats.

Explain that missed runs become one catch-up run before the usual schedule
continues, disabled tasks stay off, and tasks interrupted while running still
need attention. Update the plugin contract and static coverage for the card's
position, copy, and screenshot.

Validation: 10 focused tests passed before the title clarification; all six
static tests passed afterward. Verified navigation to card 2 and captured it
in the live WebUI before the title clarification.
2026-11-17 15:07:08 +01:00
Alessandro Frau
aa02115056 Add A0Cleaner to What's New for existing users
Some checks are pending
Build And Publish Docker Images / plan (push) Waiting to run
Build And Publish Docker Images / build (push) Blocked by required conditions
Introduce the A0Cleaner card with a screenshot, cleanup schedule and manual
review guidance, and the plugin entry point. Render bullet headings in bold
while keeping all card copy escaped through x-text.

Check model readiness before automatic display. New users mark the current
release as seen so completing onboarding or reloading does not show its
highlights; future releases remain eligible. Failed readiness checks skip
display without recording the release as seen.

Update the plugin contract and add regression coverage for first-run
suppression, configured users, future updates, and readiness failures.

Validation: 10 focused tests passed across tests/test_whats_new.mjs and
tests/test_whats_new_static.py. Live browser checks verified startup,
dismissal, and first-run suppression before the final copy revision.
2026-11-17 14:45:29 +01:00
Alessandro Frau
be1e3afca2 Guide new users to model setup in the welcome composer
Some checks are pending
Build And Publish Docker Images / plan (push) Waiting to run
Build And Publish Docker Images / build (push) Blocked by required conditions
Show "Ask anything to set up your models, or click Settings" when Welcome reports missing API credentials. Reuse the existing banner state so the hint updates after setup without adding a readiness request.

Preserve the usual welcome, chat and queue placeholders. Add regression coverage for missing credentials and the return to the normal placeholder, and document the behavior.

Validation: 20 focused tests passed. Visual verification was blocked by browser-tool errors; the full suite was not run.
2026-11-16 19:37:55 +01:00
Alessandro Frau
2ced8b38a6 Require authentication for OAuth model proxies
Replace loopback trust with an internal process credential supplied automatically to local account-model calls. Keep configured external proxy tokens and deny tokenless requests from every source address.

Finalize local proxy credentials after model defaults are merged so legacy stored keys, inline keys, additional parameters and auth headers work without changing saved settings or account connections. Limit replacement to this process's loopback WebUI port and provider proxy path, preserving explicit credentials for other endpoints.

Keep the internal credential in the stable framework runtime module so plugin reloads do not invalidate existing model clients. No saved credential or configuration migration is required.

Isolate proxy tests from saved accounts and plugin toggles. Cover all account providers, disconnected state, custom ports, external endpoint credentials, public hostnames, forwarded headers, OAuth callback accessibility and real SDK HTTP streaming. Document the intentional requirement for external clients that previously relied on tokenless loopback trust.

Validation: 643 focused framework tests passed with 3 expected failures for two pre-existing backup/A2A issues, plus 52 isolated host checks and the OAuth UI regression. A live Codex Responses call using the existing account after plugin reload returned OK; saved credentials and configuration remained byte-for-byte unchanged.
2026-11-16 19:11:08 +01:00
Alessandro Frau
56b2910f46 Stop A2A ambient token forwarding
Do not automatically attach the process-wide A2A_TOKEN to model-selected destinations. Keep explicitly supplied connection credentials and document the existing token-bearing Agent Zero connection URL for protected peers.

Add regressions proving that arbitrary destinations receive no ambient Authorization or X-API-KEY headers, explicit tokens still work, and token-bearing URLs are preserved for discovery and RPC. Exercise discovery, message send and conversation reuse through the server auth/router for token URLs, explicit headers and public peers.

Validation: 8 client/tool tests passed. One expected failure records pre-existing project URL handling of chunked request bodies; the server implementation is unchanged.
2026-11-16 19:11:08 +01:00
Alessandro Frau
5b315acfae Fix backup restore path traversal
Normalize logical archive targets and reject targets whose resolved paths escape the Agent Zero restore root. Validate every member before cleanup or extraction, then recheck each target before use.

Preserve symlink names for restore patterns, overwrite policies and destination ALLOWED_ORIGINS handling. Cover traversal and symlink escapes without changing existing files, contained symlinks, symlinked installation roots and environment files.

Exercise portable backup creation, inspection, preview and restore across different VPS installation paths, with saved OAuth credentials, model configuration, cleanup and overwrite policies.

Validation: 27 backup tests passed in the framework runtime. Two expected failures record the pre-existing clean-plus-skip preview count mismatch after verifying restored data.
2026-11-16 19:11:08 +01:00
Alessandro
728d67a0cb Fix A0Cleaner toast dismissal when opening cleanup
Dismiss the notification toast and mark it as read when Clean up opens the A0Cleaner modal. Reuse the notification store's existing dismissal helper, add a regression assertion, and document the behavior.

Validation: 34 focused Python tests and 6 cleaner store tests passed. Verified on localhost:32081 that the modal opens, the toast disappears, and the notification is marked read without executing cleanup.
2026-11-16 16:11:32 +01:00
Alessandro
d3cb2323a0 Document scheduled task creation through the UI and chat
Some checks failed
Build And Publish Docker Images / plan (push) Has been cancelled
Build And Publish Docker Images / build (push) Has been cancelled
Add a complete Scheduled Tasks guide with real form and chat examples, six framed screenshots, schedule fields, task conversations, controls, and catch-up behavior. Link it from the docs index, Usage, and A0Cleaner, and correct the Cleaner timezone instructions.

Verified both creation paths on localhost:32081, ran the form-created task, checked edit/reset/delete controls, and removed the demo tasks. Validated local links, matching website text, image dimensions, and rendered screenshots. Document the current lack of a Stop button for scheduler runs.
2026-11-14 22:03:58 +01:00
Alessandro
7027b969f6 Document A0Cleaner workflows with framed screenshots
Add a practical guide to opening A0Cleaner, automatic cleanup defaults,
notification timing, missed schedules, manual review, custom folders, and
cleanup history. Explain the difference between one-time selection and Auto
switches, and clarify the limits around Linux /tmp and uploaded files.

Include five real UI captures with complete panel framing and neutral
surroundings. Reproduce the native notification from the saved maintenance
result and leave the custom-folder example unsaved. Optimize the PNG set
from 311 KB to 119 KB. Link the guide from the docs index and Usage, and
record ownership alongside the aligned website guide.

Validation:
- Verified the entry point, defaults, and review/cancel flow at localhost:32081.
- Checked 33 local links, anchors, components, and assets across both guides.
- Confirmed matching guide text and no em dashes.
- Inspected optimized screenshots and the rendered Core documentation preview.
- git diff --cached --check passed.
2026-11-14 21:00:41 +01:00
Alessandro
3d2b2eaa2a Remove A0Cleaner from the sidebar dropdown
Keep access through plugin settings and the cleanup notification, and update the owning documentation to match. Leave Linux /tmp and uploaded attachments outside built-in cleanup targets.
2026-11-14 20:41:23 +01:00
Alessandro
2a711f40a2 Reuse filesystem watches to reduce backend startup time
Retain observer subscriptions for unchanged covering roots instead of tearing
all watches down whenever a logical subscription is registered or removed.
Preserve event filters, callbacks, debouncing, and recursive coverage while
tracking successful subscriptions so failed registrations can be retried.

Reuse frdel's refresh-lock pattern from 188cff20 to keep observer scheduling
and shutdown outside the registry lock and drain concurrent registrations
without inverting the observer and callback lock order.

On the Windows checkout mounted into WSL, plugin-tree scans dropped from five
to one. Instrumented startup fell from 95.5s to 46.1s; ordinary supervised
restarts took 49-51s. No watchers or startup features were disabled.

Validation: 128 focused Linux tests and six native Windows watchdog tests
passed, including real filesystem event delivery. Update the owning helper
and test contracts alongside the implementation.
2026-11-14 18:11:42 +01:00
Alessandro
b1887b9a73 Reduce filesystem overhead when building WebUI asset bundles
Reuse scandir metadata and resolve only file symlinks during inventory traversal. Reuse verified extension paths while retaining full resolution for requested assets, deterministic bundle ordering, and containment checks. Skip broken links and unreadable directories.

On a Windows-hosted checkout mounted into WSL Docker, profiled bundle preparation fell from 39.7s to 17.7s and first UI opening from 22.2s to 19.1s. Cached openings took about 3-4s.

Validation: 13 focused bundler, startup-assets, and service-worker tests passed in the framework runtime, including symlink escape, loop, and unreadable-directory regressions. The live Instance rebuilt all 333 bundled assets after restart.
2026-11-14 17:43:58 +01:00
Alessandro
9048880222 Catch up recurring scheduler tasks after downtime
Some checks failed
Build And Publish Docker Images / plan (push) Has been cancelled
Build And Publish Docker Images / build (push) Has been cancelled
Persist the schedule anchor and run overdue idle tasks once on the next tick, coalescing missed occurrences instead of relying on a one-minute lookback. Preserve pending work across ordinary edits, reset timing when schedules or states change, and carry the anchor through JSON, YAML and public serializers.

Recheck eligibility when claiming a run and keep one worker per task to prevent stale selections and repeated ticks from launching duplicates. Calculate cron delays in the task timezone across daylight-saving changes. Document disabled, failed and interrupted-task behavior in the scheduler and its skill.

Validate with 195 focused regression tests. In the live Docker stop/start test, the enabled marker-only task ran exactly once 93 seconds after its missed slot; the disabled task remained untouched, two extra ticks caused no replay, and existing tasks were unchanged.
2026-11-13 12:07:03 +01:00
Alessandro
c641619a77 Add A0Cleaner with automatic bulk cleanup and weekly maintenance
Clean aged download caches, installer leftovers, retired browser profiles, inactive-chat screenshots, temporary captures and unused Chromium versions by default. Preserve explicit master and per-area opt-outs; keep embedding caches and custom folders behind manual selection and confirmation.

Add bounded metadata scans, identity-checked deletion, active-resource protection, persistent activity, editable native scheduler integration and actionable A0 notifications. Provide a responsive branded UI and a 256px thumbnail based on the Agent Zero logo.

Validate with 65 focused backend tests, 6 frontend tests and live desktop/mobile checks on localhost:32081. The native task cleaned 88,833 files (4.4 GB) in about 22 seconds; protected chat text, uploads, shared browser data, required Chromium and embeddings were preserved.
2026-11-13 11:24:18 +01:00
Alessandro
de70d23b04 Wait for task detail to close before opening its editor
Close the scheduler detail modal by its explicit path and await its asynchronous close hooks before opening the editor. This prevents the new editor from being removed from the modal stack and leaving an uncloseable overlay. Cover delayed closure in the editor regression test and verify the transition on the live instance.
2026-11-13 11:22:18 +01:00
Alessandro
e5475447de Preserve cancellation when a maintenance hook stops with an error
Wait for the cooperative plugin worker to stop, then retain the original scheduler cancellation even when the hook raises during shutdown. Cover both returning and raising hooks so cancelled maintenance does not become a failed task.
2026-11-13 11:14:35 +01:00
Alessandro
60b564bc57 Run plugin maintenance through the A0 task scheduler
Add an optional plugin owner to saved tasks and execute its scheduled hook without creating an agent chat or calling a model. Keep schedule, state, cancellation and persisted results in the existing scheduler; disabled or missing handlers fail closed.

Hide inapplicable model inputs for plugin tasks and preserve legacy task behavior. Verified with 37 scheduler execution, recovery, serialization, timezone and editor tests.
2026-11-13 10:15:33 +01:00
Alessandro
cafcfc9545 Release idle Browser resources without interrupting agents
Some checks are pending
Build And Publish Docker Images / build (push) Blocked by required conditions
Build And Publish Docker Images / plan (push) Waiting to run
The shared Chromium, Xvfb and Xpra runtime could remain alive indefinitely
after chats stopped using it. Suspend it after one hour of inactivity by
default, preserving saved tab URLs and the shared sign-in profile. Protect
running and paused owners, subordinate parent tasks, connected viewers,
browser calls and downloads; passive polling must not restart the runtime
or extend its lifetime.

Add Stop internal browser to Browser Settings, release the runtime on
global plugin disable, and clean stale display directories only when no
live owner or process reference remains. Document the global timeout,
opt-out and loss of unsaved page state on suspension.

Validation: 206 focused tests and 20 real-Chromium security tests passed.
Live Docker checks verified periodic and manual shutdown from 13 child
processes to zero, download protection, tab/cookie recovery, the WebUI
control and CSRF rejection. The reported GPU busy-loop was not reproduced;
this change bounds idle lifetime without changing graphics flags.

Refs #1892, #2029.
2026-11-12 20:59:02 +01:00
Alessandro
04e9cf29f0 Document Telegram setup, scopes, and upgrade recovery
Add a practical Telegram guide covering BotFather setup, private polling,
access and administrator lists, commands, attachments, project/profile
ownership, groups, webhooks, and preservation of existing configurations.
Explain distinct bot accounts per saved scope, inheritance, duplicate-token
recovery, and the limits of connection checks and default project routing.

Link the guide from the main README, documentation index, and plugin
reference. Add four optimized screenshots using neutral unsaved drafts,
with complete dialog frames and comfortable spacing. Record that screenshot
framing requirement in the documentation contracts.

Validation:
- Checked guide links, anchors, and assets alongside the website counterpart.
- Verified UI labels and screenshots against the live Docker instance.
- Confirmed Telegram configuration and activation files stayed unchanged.
- git diff --check passed.
2026-11-12 18:00:15 +01:00
Alessandro
e4d1905fb3 Clarify default remote input access
Clarify that input_remote and remote terminal execution are already enabled
by default and require no Agent Editor grants. Update the tool prompt,
host execution skill, and policy documentation to distinguish this default
from custom profiles that restrict terminal access.

Add regressions covering prompt and native-schema visibility, policy
checks, and dispatch without explicit grants under inherited and custom
allow-by-default policies. Keep terminal dependency enforcement and host
execution and Read&Write gates unchanged.

Validation:
- 192 focused tests passed in both Conda a0 and Docker's framework runtime.
- Live runtime policy checks confirmed input, input_remote, and remote
  execution are allowed for the default and agent0 profiles.
- No new end-to-end CLI session was run.
2026-11-12 17:39:18 +01:00
Alessandro
9ae1939bd6 Restore secure Telegram scopes and preserve legacy configuration
Some checks are pending
Build And Publish Docker Images / plan (push) Waiting to run
Build And Publish Docker Images / build (push) Blocked by required conditions
Restore global, project, profile, and project/profile configuration with
one bot identity per saved scope. Preserve inherited configuration and
pause conflicting legacy bot identities without discarding their settings.

Revalidate scope, activation, and access rules for inbound messages,
commands, callbacks, session selection, and outbound delivery. Keep scoped
bots within their owning project/profile, omit credentials from persisted
chat metadata, and sanitize attachment filenames.

Automatically back up existing configuration, activation, and session
mappings before upgrade. Preserve their contents and stable global session
keys, while provisioning secure webhook secrets where legacy settings
require them. Document scope ownership, conflict recovery, and migration.

Validation:
- Docker framework suite: 216 passed, plus 73 subtests.
- Live Telegram status and message/reply round-trip confirmed.
- Missing and forged webhook credentials rejected with HTTP 403.
- Existing live configuration, activation, and session mappings unchanged.
2026-11-12 17:27:15 +01:00
Alessandro
c84ff9521a Enable Time Travel snapshot expiry by default
Default active-history expiry to on in bundled settings, backend fallbacks
and the settings UI. New installations and legacy configurations without
the key now keep up to 500 recent snapshots within 30 days, plus current
state and boundary checkpoints. Preserve explicit opt-outs and the existing
cleanup schedule.

Update the settings copy and ownership contract to distinguish snapshot
expiry from optional whole-history deletion for inactive workspaces.

Validation: 59 focused backend tests and 5 frontend tests pass. Verify
that default cleanup expires old preserved timelines, explicit opt-outs
leave refs unchanged, and the settings initializer preserves false values.
The Docker framework runtime reports expiry enabled with the 500/30 limits.
2026-11-12 16:14:42 +01:00
Alessandro
5fc8826a41 Add opt-in Time Travel retention and capture controls
Add workspace-relative exclusions and an optional file-size limit, report
skipped files and history storage, and protect excluded files during restore.
Fail incomplete scans and stop queued automatic captures when disabled.

Keep snapshot expiry off by default. When enabled, retain recent snapshots,
the current state and full-tree checkpoint boundaries; replace refs atomically
and reclaim expired objects under the workspace lock. Retry interrupted GC,
stop child processes on maintenance timeout, and preserve workspace inactivity
across compaction. Checkpoints remain restorable but cannot be reverted.

Keep existing saved settings compatible and expose coverage and retention
controls in the plugin UI. Review against relevant frdel history informed
shared ownership and consistent capture/restore filtering.

Validation: 57 focused backend tests in the Docker framework runtime and 5
frontend store tests pass. Live UI checks covered workspace selection, diffs,
Travel, Revert, explicit current inspection, checkpoint actions, and legacy
settings save/reopen. Original settings were restored and disposable QA
project/history removed.
2026-11-12 16:11:15 +01:00
Alessandro
28e36b5fef Load Time Travel history without staging workspace files
Page saved commits directly, batch metadata and file statistics, and bound file-filter scans to Git batches. Keep the snapshot index stat cache instead of clearing and rebuilding it on every capture.

Inspect current changes explicitly in a disposable repository borrowing saved objects read-only. Offload Git work from API event loops and discard stale UI responses when the selection or workspace changes.

Verified with 42 Python tests and 3 frontend tests, including unchanged saved-repository contents after inspection, unusual filenames, filtering, and a 25-process ceiling for a ten-entry history page.
2026-11-12 15:41:23 +01:00
Alessandro
78d3f8d27d Protect Time Travel capture, restores, and cleanup
Treat Git paths literally so unusual filenames cannot expand into excluded files. Reject restores that would change excluded paths, nested Git workspaces, or files through directory symlinks.

Serialize complete workspace operations with reentrant cross-process locks shared by retention. Skip busy or symlinked histories, fail closed when owners cannot be enumerated, and count only successful cleanup.

Verified with 38 focused tests, including cross-process cleanup, concurrent snapshots, exclusion bypasses, and restore safety. Keep test shadow repositories isolated from user history.
2026-11-12 15:35:38 +01:00
Alessandro
797a88faec Keep welcome sends visible while initial chat history loads
Some checks failed
Build And Publish Docker Images / plan (push) Has been cancelled
Build And Publish Docker Images / build (push) Has been cancelled
The first history snapshot could clear the optimistic user bubble after a
Welcome send created and selected a chat, leaving the prompt invisible
until its backend log arrived.

Skip the redundant history reset on the first snapshot and rebuild the
merged message window in log order. Reuse the existing message IDs and
renderer so the greeting precedes the prompt and the backend echo updates
the same user row. Later full snapshots still clear stale history.

Add browser regressions for text and attachment sends with snapshots
arriving during or after rendering, and update the WebUI contracts.

Validation: 72 focused tests passed. Live desktop and mobile checks covered
delayed dispatch, normal in-chat sending, and a real backend message echo.
2026-11-10 19:30:54 +01:00
Alessandro
e824cf5d99 Fix chat loading for malformed code execution arguments
Reuse the existing command text conversion before building execution
titles so rejected non-string arguments no longer abort streamed
rendering or saved-chat replay. Preserve original log arguments and
normal command copying and output.

Add regressions for malformed arguments in both delivery paths and
document the renderer contract.

Validation: 71 focused tests passed; both previously affected DeepSeek
chats reopened and reloaded with zero browser console errors.
2026-11-10 17:57:51 +01:00
Alessandro Frau
70d07d8667
Merge pull request #2004 from DZH166/codex/fix-code-execution-code-args-20261102
fix(code-execution): reject non-string code before shell dispatch
2026-11-10 17:27:00 +01:00
Alessandro Frau
ec09f65d23
Merge pull request #1992 from tanishqbaweja/fix/1773-dirty-json-surrogates
fix: decode escaped surrogate pairs in DirtyJson
2026-11-10 16:54:23 +01:00
Alessandro Frau
4fb17f9512
Merge pull request #1965 from MohammadHijjawi97/fix-text-editor-append-unterminated-last-line
Keep line-edit appends off an unterminated last line
2026-11-10 16:54:09 +01:00
Alessandro
6bd26ffe7c fix(secrets): refresh cached credentials after dotenv writes
Some checks are pending
Build And Publish Docker Images / plan (push) Waiting to run
Build And Publish Docker Images / build (push) Blocked by required conditions
Invalidate registered secrets managers after every successful central dotenv write, including writes that skip environment reload. Import the manager lazily to preserve the existing dotenv/secrets dependency cycle.

Cover setting, rotating, and clearing credentials through direct writes and the model configuration backend so substitution and masking see current values.
2026-11-10 14:06:54 +01:00
Alessandro
2fca75d69b fix(secrets): mask provider key pools and split stream chunks
Retain possible secret prefixes from the first streamed character while preserving the existing final partial-mask threshold.

Mask full provider key lists and each selected member in normal and streaming output, including supported legacy key names. Preserve stored values, placeholder substitution, text mask thresholds, and commas in ordinary secrets. Cover chunk boundaries, pool whitespace, and naming variants.
2026-11-10 14:06:54 +01:00
Alessandro
01e0ced1be fix(secrets): preserve values across edits and read failures
Use dotenv parsing to distinguish comments from secret content, mask shadowed populated definitions, and resolve editor masks against current disk values instead of stale snapshots.

Treat missing stores as empty, reject unreadable-store edits, and retain an unavailable value through settings normalization so unrelated updates cannot erase secrets, even after a transient read failure recovers. Cover failed reads, recovery, rotation, and masked round trips.
2026-11-10 14:06:49 +01:00
Alessandro
33e29a3b2e fix(secrets): omit empty and overridden prompt aliases
List only non-empty effective definitions, honoring case-insensitive last-definition precedence across global, native, and project stores. Exclude malformed source lines from the alias inventory.

Keep the existing conditional secrets guidance and independent variables section. Cover empty values, overrides, non-empty literals, and rendered prompt behavior.
2026-11-10 14:06:45 +01:00
Alessandro
dc64354cc9 ui: surface unusable response limit in agent settings
Some checks failed
Build And Publish Docker Images / plan (push) Has been cancelled
Build And Publish Docker Images / build (push) Has been cancelled
Place the response limit directly below the default agent profile. Keep knowledge subdirectory and project inheritance in Advanced Settings.
2026-11-04 15:26:11 +01:00
DZH166
5fb00bd909 fix(code-execution): reject non-string code before shell dispatch 2026-11-03 00:29:20 +08:00
Alessandro
ba9b67df0f docs: refresh Launcher screenshot in README
Some checks failed
Build And Publish Docker Images / plan (push) Has been cancelled
Build And Publish Docker Images / build (push) Has been cancelled
Replace the compact Runtime Ready capture with a full-width installation
walkthrough screenshot. Preserve the complete scene and resolution, remove
the superseded asset, and update the image description and ownership docs.

Optimize the PNG from 324,982 to 122,479 bytes. The Launcher repository uses
an identical copy.

Validation: checked image paths and placement, PNG decoding, matching asset
hashes, local desktop/mobile rendering, and git diff --check.
2026-11-01 22:10:32 +01:00
Alessandro
cb1845217b docs: showcase Launcher onboarding and prioritize Apple Silicon
Add the Runtime Ready onboarding screenshot directly below Desktop setup.
Present recommended downloads by platform, with Apple Silicon first for
macOS and the alternative clearly labeled Intel Macs only. Keep processor
labels concise and preserve all existing download URLs.

Update the documentation contract for the screenshot and Mac download order.

Validation: checked links, screenshot loading, Apple Silicon placement,
local desktop/mobile rendering, and git diff --check.
2026-11-01 22:03:41 +01:00
Alessandro
f969c5674c docs: clarify desktop and shell installation options
Present Launcher downloads by platform and processor, and recommend A0
Install in a visible Servers and shell access section. Share first-message
setup across both paths and keep direct Docker and troubleshooting under
Other install options.

Link scripted setup to the installation guide, preserve existing anchors
and download URLs, and update the documentation contract.

Validation: checked 19 Quick Start links and anchors, local desktop/mobile
rendering, accordion behavior, unchanged Docker command, and git diff --check.
2026-11-01 21:58:04 +01:00
Alessandro
ec87319db6 docs: refresh README Plugin Hub screenshot
Replace the Plugin Hub screenshot with the updated community plugin browser capture. Optimize the PNG while preserving its dimensions and existing README reference.
2026-11-01 21:39:05 +01:00
Alessandro
dd23dc0e33 docs: fix broken DeepWiki badge in README
Replace the DeepWiki-hosted badge image with a Shields.io badge matching the other README buttons. Preserve the link to the Agent Zero DeepWiki page.

Validation: confirmed HTTP 200 and valid SVG for the replacement badge; git diff --check passed.
2026-11-01 21:24:29 +01:00
Alessandro
dbf6849c84 fix(scheduler): use Scheduled Tasks in modal and navigation
Some checks are pending
Build And Publish Docker Images / plan (push) Waiting to run
Build And Publish Docker Images / build (push) Blocked by required conditions
Align the scheduler modal title and sidebar dropdown label with the welcome screen Scheduled Tasks card. Update the scheduler documentation and plugin index to match.

Validation: 97 focused scheduler, timezone, tool contract, and WebUI extension tests passed. Confirmed the live instance serves the updated labels.
2026-11-01 20:13:55 +01:00
Alessandro
fcf4a2c599 Fix welcome panel control visibility and ordering
Apply Quick Actions hover and focus visibility to all header controls,
including resize, while preserving touch access and open move menus.
Place the Quick Actions and System Resources drag handles immediately
before their dismiss buttons and document the shared ordering rule.

Validation: 11 welcome Node tests and 6 composer pytest tests passed.
Verified desktop hover, keyboard focus and tab order, width toggles,
move-menu behavior, and mobile touch visibility against localhost:32081.
2026-11-01 13:08:34 +01:00
Alessandro
5f45daa61a Reject unsupported Telegram configuration scopes
Telegram advertised project/profile settings while its lifecycle and sender
authorization read only global configuration. A saved scoped allowlist or
disable could therefore leave a permissive global bot accepting messages.

Make Telegram configuration and activation explicitly global, preserving
per-user project routing. Enforce manifest scope support before plugin
configuration or activation writes and return HTTP 400 for invalid saves.

Pause Telegram bots and block incoming events whenever legacy scoped
configurations or activation overrides exist. Preserve those files, skip
attachment cleanup, notify the operator, and retain settings UI access for
manual review and removal. Each event must resolve a currently enabled
global bot; removed, renamed, disabled, or globally deactivated bots cannot
fall back to captured settings.

Add regressions for scope rejection without side effects, supported-plugin
scopes, legacy settings shutdown and migration, current allowlists, and
immediate revocation. Update existing administrator tests and owning docs.

Validation: 339 tests and 73 subtests passed in both Conda a0 and Docker's
framework runtime. A before/after reproduction showed an excluded sender
reaching context creation before the fix and being blocked afterward.
Restarted the live runtime and checked HTTP health and Telegram settings UI.
No live Telegram message test was performed.
2026-11-01 08:05:25 +01:00
Alessandro
25542e9950 Require remote terminal permission for input_remote
input_remote forwards keyboard text as terminal code, allowing a custom
profile to bypass a blocked code_execution_remote capability through an
independently enabled input tool.

Require the canonical bundled remote terminal action in shared policy
resolution while retaining input_remote's independent permission. Apply
the dependency to execution, text prompts, and native tool schemas, and
recheck it inside the input wrapper before rewriting args or dispatching.
Output-only grants, default-deny policies, and profile tool-name overrides
cannot authorize the terminal operation.

Document the permission dependency and add 19 regressions covering zero
dispatch on denial, action-specific grants, prompt/schema visibility,
authorized input, host access controls, and framework-owned allow_running.

Validation: 232 focused tests passed in both Conda a0 and Docker's framework
runtime. A smoke test through the real CLI shell backend verified denied
dispatch, authorized idle and interactive input, and permission revocation.
Restarted the live Agent Zero runtime and verified HTTP health.
2026-11-01 07:49:29 +01:00
Alessandro
c6b3f56d72 Enforce same-origin isolation for browser DOM helper RPC
Cross-origin related windows could read page content and invoke DOM actions
through the helper's public postMessage channel.

Reject foreign and opaque origins, use exact target origins, and bind each
pending response to its child window and operation. Preserve inherited
origins for about:blank/srcdoc frames and trusted direct automation access.

Bump the helper to v2 and require a page reload when legacy listeners are
present. Existing host browsers must restart and reconnect before restored
tabs reload so old initialization scripts are discarded.

Add real-Chromium regressions for parent/frame/opener attacks, opaque
origins, forged responses, same-origin actions, and helper upgrades.

Validation: 250 browser/connector tests passed in both Conda a0 and Docker's
framework runtime; JavaScript syntax and diff checks passed. Restarted the
live runtime and verified HTTP health, browser navigation, and the viewer.
2026-11-01 07:41:16 +01:00
Alessandro
4766b114ae Require explicit WhatsApp administrators for settings changes
Accepted WhatsApp senders could persistently relax their own tool policy
and toggle instance-wide plugins using ordinary chat authorization.

Add a separate admin_numbers allowlist, empty by default, and derive
administrative access from the normalized sender of each inbound message.
Keep allowed-number and group checks in force. Require administration at
the WhatsApp plugin-toggle and tool-policy mutation boundaries while
preserving read-only menus and ordinary chat controls.

Gate built-in profile creation and browser-runtime changes before shared
command execution. Use the shared argument parser for profile creation
checks and preserve existing profile selection and custom overrides.
Expose the administrator list in settings and document the new defaults.

Add regressions for denied writes, real policy persistence and runtime
decisions, normalized sender identities, revocation, administrative
commands, and independent frontend access-list editing.

Validation:
- Conda framework: 165 tests and 62 subtests passed.
- Docker framework: 12 focused tests and 32 subtests passed.
- Node bridge and configuration store: all 4 tests passed.
- WebUI: administrator field renders, parses and clears independently.
- JavaScript syntax and git diff checks passed.
- Restarted the bound Docker instance and verified HTTP 200.

Live WhatsApp delivery was not exercised. Synthetic events and disposable
policy files cover the authorization paths without changing live access
lists or sending test messages to WhatsApp users.
2026-11-01 07:28:02 +01:00
Alessandro
1eba072653 Require explicit Telegram administrators for settings changes
Some checks are pending
Build And Publish Docker Images / plan (push) Waiting to run
Build And Publish Docker Images / build (push) Blocked by required conditions
Accepted Telegram senders could use /permissions and its callbacks to
persistently remove their own tool restrictions. The same access also
exposed instance-wide plugin toggles.

Add a per-bot admin_users allowlist and enforce it when opening settings
and processing every settings callback. Administrators must still pass
allowed_users; an omitted or empty admin_users list grants nobody settings
access. Resolve membership from the current bot configuration on each
event so revocation applies to existing buttons.

Expose the administrator list in the WebUI and document its defaults.
Add regression coverage for message and callback routes, denied writes,
authorized policy persistence, runtime tool decisions, and revocation.

Validation:
- Conda framework: 123 tests and 41 subtests passed.
- Docker framework: 9 focused tests and 41 subtests passed.
- WebUI: administrator field renders, parses entries, and clears correctly.
- JavaScript syntax and git diff checks passed.
- Restarted the bound Docker instance and verified HTTP 200.

Telegram network delivery was not exercised; tests use synthetic events
and disposable policy files without changing live bot access settings.
2026-11-01 07:22:49 +01:00
Alessandro
c7c733ec04 Copy full execution commands while keeping compact previews
Restore matching command previews from the stored, masked command when
copying execution output or opening step details. Preserve command
whitespace and the output body while keeping the chat preview compact
and retaining existing log size limits.

Cover Bash, PowerShell, Python and Node.js prefixes, streamed updates,
masked placeholders and output without a matching command preview.

Validation: 48 focused WebUI tests passed, git diff --check passed, and
an actual browser clipboard check passed using frontend assets served
by the live instance at localhost:32081.
2026-11-01 07:16:55 +01:00