- GET /scenarios/:id/export-e2e and the export_e2e_test MCP tool take
an optional version, defaulting to the latest
- build the test project from that version's steps, name and timeout
- put the version in the zip name (<scenario>-vN-e2e-test.zip), the
generated package name and the README heading
- the Export e2e test button exports the version being viewed
- document the e2e export in scenario.md and mcp.md
- reopen the run events stream with backoff when the browser closes it for good (Firefox drops quiet connections after a network change)
- enable notifications when the permission is granted outside the bell, unless turned off explicitly
- follow permission changes made in the browser's site settings
- highlight the bell while notifications are on
- add RunEventsService and emit run.started / run.finished from the scheduler
- stream run events over SSE at GET /api/v1/scenarios/runs/events with a heartbeat
- add useRunNotifications hook and NotificationToggle (opt-in bell in the side panel footer)
- refresh runs lists when a run finishes
- document the events stream and notifications
Closes#12
- request the exact version being exported via ?version=N so file
content always matches the filename
- name downloads <scenario>-v<N>.yaml and snippet-<alias>-v<N>.yaml
- read answers from /dev/tty: under `curl ... | bash` stdin is the
script itself, so `read` consumed the script's own lines; fall back
to defaults when there is no terminal
- check that docker, the compose plugin and the daemon work and that
compose accepts the generated compose.yaml
- start liqa with `docker compose up -d` in the install directory,
printing the manual start command if a check or the start fails
- examples/self-test: importable scenario, snippets, credential and
environment that test the liqa instance they run on
- CI builds the Docker image only for vX.Y.Z release tags
- bump version to 1.24.3
- add examples/ with a ready-to-import self-test: environment, credential,
snippets and a scenario that creates, runs and checks another scenario
through the liqa REST API and UI, then cleans up
- add test/examples.spec.ts: imports every example and parses its code
- link the examples from README, docs/scenario.md and CLAUDE.md
Closes#6
- trigger the image job on vX.Y.Z tag pushes instead of master pushes
- run server and client checks unconditionally on tags before the build
- fail the build when the tag doesn't match the package.json version
- document the release routine: version bump, changelog and tag are
done on master after work branches are merged
- large run log data blocks start collapsed, with per-block and
expand/collapse-all toggles; limits set by RUN_LOG_DATA_COLLAPSE_BYTES
and RUN_LOG_DATA_COLLAPSE_LINES
- copy logs button copies the filtered logs as plain text
- GET /manifest.json serves the app version and public client settings,
cached by the UI in localStorage
- bump version to 1.24.2
- collapse run log data blocks whose pretty-printed JSON exceeds a byte
or line limit, with a per-block toggle and an expand/collapse-all
button next to the Logs heading
- align data blocks and toggles with the context/message column
- add a "Copy logs" button that copies the filtered logs as plain text
with every data block expanded
- add GET /manifest.json with the app name, version and run log data
limits (RUN_LOG_DATA_COLLAPSE_BYTES / RUN_LOG_DATA_COLLAPSE_LINES)
- load the manifest in the client stale-while-revalidate: localStorage
copy first, refetched per page load and on tab focus after a while
- show the server version in the sidebar footer, with a reload hint
when it differs from the client build
Closes#3
- "Step N passed" log entries carry the step's return value as data
- context.assert takes an optional details argument stored as log data
on failure; step and snippet failure entries name only the assertion
- log output shows elapsed time in whole seconds
- run artifact and scenario file names open the file in a new tab
- bump version to 1.24.1
- "Step N passed" log entry carries the step's return value as data
- context.assert accepts an optional details argument; on failure the
assertion:fail entry stores { details, error } as data
- assertion failures throw AssertionFailedError so "Step N failed" and
"Snippet X failed" entries name only the assertion, without repeating
the details; the step run description keeps the full message
- exported e2e runtime accepts and prints assert details
- log output shows elapsed time in whole seconds, matching the stored
timestamp resolution
- run artifact and scenario file names link to the inline file content
in a new tab
- step and snippet editors open a reference of the script runtime and
its context API from a (?) button next to the Code label
- bump version to 1.24.0
- a (?) button next to the Code label on the step and snippet
create/edit pages opens a reference of the script runtime: globals,
every context member with examples, and pitfalls such as awaiting
getStepOutput and assert
- the reference lives in components/modals/code-help.md and is
rendered as markdown
- Modal takes size="lg" for document-like content; the dialog is
capped to the viewport and its body scrolls, focusable when lg so
the keyboard can scroll it
- docs/scenario.md notes that getStepOutput is async and points to
the in-app reference to keep in sync
- scenario run log entries carry a context (step/snippet/assertion/run)
and a data payload split out from the message, plus a new debug level
- run detail page shows the full log list in a terminal-style viewer
instead of a paginated table
- bump version to 1.23.0
- context.log/warn/error and console.* split arguments: strings join
into message, object/array args are kept out of it and stored as
the log entry's new data column (JSON), so message stays a stable
string regardless of an object's shape
- added a debug log level alongside log/warn/error, including
context.debug() and console.debug() (previously folded into log)
- CSV/Markdown run log export include the data column/block
- LogOutput renders data pretty-printed in a separate block below
the message, and dims debug entries with a dedicated grey token
calibrated to the WCAG AA contrast floor
- scenario_run_logs gains a context column (step:N, snippet:alias,
assertion:pass/fail, run), tracked through a call stack in
CodeExecutorService so nested runSnippet calls tag their own and
inner log lines
- automatic step started/passed/failed entries and the Playwright
trace upload notice are tagged accordingly
- run log export (CSV/Markdown) and the logs search match context
as well as message
- add a LogOutput UI component: a terminal-style log viewer with
level color coding, elapsed-time-since-start timestamps and a
context tag, replacing the generic Table on the run detail page
- show the full log list instead of paginating it, capped to a
scrollable block
- fix migrations.spec.ts's revert test, which only undid the last
migration instead of all of them
- server and client jobs only run when their directory (or shared
root files) changed, via a new changes job using dorny/paths-filter
- the version job now checks the registry for an existing image tag
of the current package.json version instead of diffing against the
previous commit's package.json, so the image job is skipped once
that version has already been pushed
- the image job tolerates server/client being skipped rather than
requiring both to succeed
- scripts/install.sh prompts for install directory, port and version,
then writes a compose.yaml using the published registry image and
creates a data/ directory for the SQLite volume
- README documents a one-line curl | bash install using the script's
raw URL on the Gitea instance
- add a step to tag each version-bump commit as v<version>, annotated
with that version's changelog entry verbatim
- push tags with git push --follow-tags
- add the mit LICENSE and set license and author in all package.json
files, plus a license section in the readme
- generate the license notices of all shipped production dependencies
and node.js in a dockerfile stage and serve them at /licenses.txt,
linked from the sidebar footer
- keep license files when slimming node_modules in the image
- add oci labels, with the version passed in by ci
- document licensing in the development docs
- bump version to 1.22.2
- raise the jest test timeout to 30s: building the nest app in
beforeAll can exceed the 5s default when all suites start at once
- accept any run status when reading back a just-created run, since
the scheduler may already have picked it up
- use REGISTRY (git.hq.ars9.space) for login and image tags instead of
deriving it from the server url, which is the runner's registration
address and may be internal-only
- document how jobs reach gitea when the runner uses an internal address
- run server lint, type-check, tests and build, and client lint, build
and storybook tests on master and pull requests
- when a push to master changes the package.json version, build the
docker image and push it to the gitea registry tagged with the
version and latest
- document the workflow and the REGISTRY_TOKEN secret
- install production dependencies and chromium in a separate stage
- drop source maps, typescript sources and declarations, docs,
typeorm's browser build, better-sqlite3's c sources and ffmpeg
- base the runtime on debian:bookworm-slim with only the node binary
- document the image layout
- bump version to 1.22.1
- list the snippets a scenario uses in the export dialog, checked by
default, with the versions the scenario uses
- export each selected snippet once per used version (latest and
pinned) via the new snippetIds query parameter
- resolve used snippets from the exported version's steps
- show the scenario version in the export dialog
- document the export query parameters
- bump version to 1.22.0
- add a RevealableCodeBlock ui component that renders its code only
after the user clicks Reveal, with a storybook story
- use it for the data section of the credential page
- bump version to 1.21.1
- show a circle colored by the first 6 hex digits of the uuid,
bordered with the badge text color for contrast in both themes
- lighten the uuid badge background on hover
- let Badge accept a className
- add a storybook story with six sample uuids and a randomize button
- bump version to 1.21.0
- add --color-primary-text for lime used as text or icon color
- stop dimming the idle refresh label with opacity
- darken the light danger button and lighten the dark danger hover
- define syntax highlighting colors per theme as --hl-* tokens
- darken light muted text slightly
- add a storybook contrast page, fail storybook tests on a11y violations
and show live token values in the colors story
- bump version to 1.20.2
- update client and server dependencies within their ranges, keeping
all playwright packages aligned at 1.63
- fix code flagged by react-hooks 7.1: refs written during render and
synchronous setState in effects
- add stories for the ui components that had none, with interaction
tests, and cover all DescriptionList layouts
- drop the unused mdx stories pattern
- run the storybook tests as part of the pre-commit checks
- bump version to 1.20.1
- replace synchronize with migrations, starting with a generated initial
migration that adopts databases created by synchronize
- apply pending migrations in the docker container before the server,
and refuse to start the server while migrations are pending
- share one entity list and data source between app, migration runner,
typeorm cli and tests; tests run the real migrations
- add migration:run, migration:show and migration:generate scripts and
a test that fails when entities and migrations differ
- exit with code 1 when startup fails
- bump version to 1.20.0
- allow the vite dev server origins in development and the compose port
origins in production; CORS_ORIGINS overrides both
- set NODE_ENV=production in the runtime docker image
- document both settings in .env.example and the development docs
- bump version to 1.19.3
- load .env in compose only when it exists; every setting has a default
- comment out every setting in .env.example, with ## for descriptions
and # VAR=default for settings to uncomment
- update README, development docs and CLAUDE.md accordingly
- rewrite architecture, development, mcp and scenario docs to match the
current code: single container, all mcp tools, runs, versioning,
timeouts, settings, client routing and dev setup
- rewrite README as a general overview without liquio-specific wording
- remove the unused KEYS_DIR setting and ./keys compose mount
- remove the unused result parameter from step code execution
- complete .env.example with the file storage settings and fix the
DB_PATH default
- choose the scenario's default environment when the run modal opens on
the scenario detail page, instead of in a load effect that raced with
loading the scenario and could fall back to the first environment
- set the run to a final status directly in the status filter test so
the scheduler can't change it mid-test
- bump version to 1.19.2
- base the docker image on node:24-slim and require node >=24 via
engines in the root package.json
- read app name and version from the workspace root package.json for
config, swagger, startup log and mcp server info; drop the server
workspace version
- bind test http servers to 127.0.0.1 so tests pass in containers
without ipv6 loopback
- bump version to 1.19.1
- serve the built client from the server via serve-static (CLIENT_DIR),
falling back to index.html outside /api, /mcp, /docs and /health
- replace server and client Dockerfiles and nginx with one root
Dockerfile and a single liqa compose service on port 13000
- use Playwright's chromium-headless-shell with only its required libs
and DejaVu/Liberation fonts instead of Debian chromium and Noto
- clear the npm cache in the runtime image
- bump version to 1.19.0
- add POST /import, an import page and the import_yaml mcp tool; items
are validated up front and imported in dependency order
- switch list page import buttons to the central import
- write scenario exports as separate --- yaml documents per entity
- give each test app its own files dir to avoid cross-suite races
- use node: prefix for node built-in imports
- bump version to 1.18.0
- replace any types with explicit types in scenario file listing,
import/export and mcp upload code
- type test server addresses and list callbacks, drop unused vars
- document why e2e export keeps the archiver import-require form
- remove unused scenario files total state in the client
- require clean lint in both packages in CLAUDE.md
- add DELETE /scenarios/:id/runs/:runId/files/:fileId and the
delete_run_file MCP tool; the stored file is removed once no scenario
or run references it
- add a delete button with confirmation to the run detail files table
- bump version to 1.17.0
- add CLAUDE.md with app overview, compose deployment details, doc
references, code quality checks and the release/commit workflow
- rename docker-compose.yaml to compose.yaml, the preferred compose
spec file name
- snapshot scenario and snippet content into version tables on every
content change; live tables keep holding the latest version
- record the scenario version on each run and copy its steps into the
run steps, so runs execute and show the version they were created with
- keep run step history when a step is deleted (FK set null)
- pin a snippet version with runSnippet('alias@N')
- add version picker to the run modal and version info across the UI
- add version list/get endpoints and MCP tools
- include version in exports and require it on import, restoring the
exact version by id + version
- record a trace for each scenario run and upload it as trace.zip
- add PLAYWRIGHT_TRACE_MODE (on, retain-on-failure, off) as the default
- add a trace mode option to the run modal, stored per run and reused
by repeat run; expose the default via GET /scenarios/run-defaults
- accept traceMode in POST /scenarios/:id/run and run_scenario MCP tool
- share one run modal component between scenario list and detail pages
- git-ignore server/data
- start a new run with the same environment and saveSession settings
and navigate to it; disabled when the run's environment was deleted
- remount the run detail page per run id so per-run state resets
- type ScenarioRun.environmentId as nullable to match the backend
- add list/get/create/update/delete_credential MCP tools so credentials
can be created and attached to scenarios entirely via MCP
- update_credential only changes the fields passed, keeping data intact
- add changelog entries for 1.13.0 and the missing 1.12.0 release
Captures a full-page screenshot and the current page HTML whenever a
scenario or snippet step fails, and uploads both as run artifacts so
failures can be diagnosed from the run's file list.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- Added E2eExportService to package scenarios as plug-and-play zip archives
containing package.json, playwright.config.ts, test.spec.ts, .env, credentials.json,
snippets, and referenced scenario files
- Only bundles credentials, snippets, and files actually referenced (directly or
transitively) by the scenario's steps, reducing archive size
- Exported test runs entirely offline using a context shim that mirrors liqa's API
(page, getCredential, runSnippet, getScenarioFiles, downloadFile, assert, etc.)
- Added GET /scenarios/:id/export-e2e HTTP endpoint and export_e2e_test MCP tool
- Added getUsedSnippets() endpoint to list snippets referenced by a scenario
- Added "Used Snippets" section on scenario detail page
- Added archiver@^7.0.1 dependency for zip archive creation
- Bumped version to 1.11.0
- Add seq column assigned in call order to fix logs racing each other
as fire-and-forget writes (e.g. a snippet's "started" entry could be
saved after its "finished" entry)
- Log snippet start/finish/failure via context.runSnippet, matching
existing step start/pass/fail logging
- Rework Markdown log export to per-entry sections instead of a table
- Version bump: 1.10.0 -> 1.10.1
- Backend: logs export endpoint for CSV and Markdown formats
- Frontend: ExportLogsModal with download button on run detail page
- Updated i18n and CHANGELOG
- Version bump: 1.9.1 -> 1.10.0
- step start, pass, and fail events logged automatically
- step boundaries visible in run logs without explicit context.log calls
- version bumped from 1.9.0 to 1.9.1
- Add context.assert(fn, description) for function assertions
- Logs outcome on success/failure, throws on assertion failure
- Bump package version from 1.8.0 to 1.9.0
Scenario steps can now persist in-script generated content (e.g. an
obtained VC's JWT) as a run file via context.downloadFile('data:...'),
without needing a network fetch. Bumps to 1.8.0.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- add scenario upload and download UI with typed file API helpers
- show run artifacts on run detail pages after polling completes
- handle multipart uploads and absolute file paths for downloads
- add FileEntity, ScenarioFileEntity, ScenarioRunFileEntity with UUID-sharded disk tree
- FileStorageService: write/sha256/expiry, paginated listing, dynamic cleanup cron via SchedulerRegistry
- expose getScenarioFiles() and downloadFile() on ScriptContext for use in exec code
- wire scenarioId, runId, fileService into ExecContextBuilder and scenario scheduler
- add file upload and listing endpoints to ScenarioController
- add FILES_DIR, FILE_RUN_EXPIRATION_DAYS, FILE_CLEANUP_CRON to AppConfig
- add usePageTitle hook that writes '<title> | Liqa' to document.title
- list pages use section name; detail pages use entity name once loaded
- create/edit pages use static action title, edit includes entity name
- register list_snippets, get_snippet, create_snippet, update_snippet, delete_snippet
- inject SnippetService into McpService and import SnippetModule into McpModule
- cover all five tools with integration tests including error cases
- extend findAll to join latest run per scenario via correlated subquery
- return lastRunStatus and lastRunAt fields in GET /scenarios response
- add Last Run column with badge and timestamp to ScenariosPage
- add integration tests for null and populated lastRunStatus
- replace sequential for-loop in pickUpPendingRuns with Effect.forEach concurrency:10
- extract CONCURRENCY_LIMIT constant shared by both startup and pickup paths
- clear all run caches (creds, snippets, env, timeouts, activeRuns) on setup failure
- call failRun on setup error to leave the db run in a consistent failed state
- use Effect.forEach with concurrency:10 to fail up to 10 runs in parallel
- isolate each failRun with catchAll so one failure doesn't abort the rest
- log unexpected failRun errors per run without halting the batch
- add onModuleInit to fail in_progress runs on server restart
(playwright sessions cannot be recovered after restart)
- extract failRun() shared method used by restart cleanup, timeout,
and unexpected error paths for consistent step run status updates
- fix catch path to also fail/cancel step runs (previously left orphaned)
- accept plain {name, steps} payload in importScenario alongside
full export format; throw BadRequestException for invalid input
- include stepRuns relation in findRuns paginated response
- add scenario-scheduler.spec.ts integration tests for onModuleInit
- add timeoutSeconds field to scenario and step entities
- add timeoutSeconds to create/update DTOs for scenario and step
- enforce timeout via scheduler: abort run if step exceeds limit
- expose timeoutSeconds in MCP create/update scenario and step tools
- add client-side type, API, i18n, and form support for timeoutSeconds
- add integration tests for timeout persistence via REST and MCP
- failStepRun was calling closeBrowserHandle immediately, removing the
handle before maybePreserveSession could read it in the finally block
- delegate browser lifecycle to maybePreserveSession for both pass and
fail paths so saveSession=true is honoured regardless of run outcome
- add sortable column support to Table with asc/desc/unsorted icons
- active sort icon uses --color-primary; no header background change
- Table supports server-side mode via total/page/onPageChange props
- wire server-side sort+pagination in ScenariosPage, AllRunsPage, RunsPage, SessionsPage
- remove steps count column from run tables
- fix server: RunsQueryDto now extends PaginationQueryDto with orderBy/orderDir
- fix findAllRuns to use QueryBuilder; sort by scenario.name via JOIN
- add per-resource typed query DTOs with @IsIn allowlist on orderBy
- prevents SQL injection and returns 400 for unknown orderBy values
- export now accepts includeEnvironment and credentialIds query params
- output is a YAML stream with comment-delimited environment, credential, and scenario sections
- scenario entity includes credentials array with alias mappings
- import accepts both old single-object and new array format, upserts envs/creds before linking
- new ExportScenarioModal with env and per-credential checkboxes replaces direct download
- add optional ManyToOne relation from scenario to environment entity
- expose environmentId in create/update DTOs, service, and MCP tools
- pre-select linked environment in run modals on detail and list pages
- add environment selector to create/edit scenario forms
- show linked environment as a navigable link on scenario detail page
- add optional description to scenario and environment entities
- expose description in create/update DTOs, MCP tools, and export DTOs
- render description as markdown on detail pages
- add description editor (CodeEditor) to create/edit forms for both resources
- move runs and sessions to top of navigation (most dynamic)
- add visual separator between dynamic and configuration sections
- reorder remaining items: scenarios, snippets, credentials, environments
- implement separator component in nav rendering logic