When ChatGPT macOS app is set to Chinese language, the "Options"
button label becomes "选项". This change checks for both English
and Chinese labels to find the button.
Co-authored-by: mad <mademing@maddeMac-mini.local>
Co-authored-by: Claude <noreply@anthropic.com>
* feat: implement Ref-Backed Locator for browser actions
Introduces a unified target resolution system with fingerprint
verification and structured error diagnostics.
Snapshot phase:
- Each interactive element now gets a fingerprint (tag, role, text,
ariaLabel, id, testId) stored in window.__opencli_ref_identity
- Zero overhead: metadata is already available during DOM walk
Resolution phase (new target-resolver.ts):
- Numeric input → ref path with fingerprint verification
- CSS-like input → querySelectorAll with uniqueness check
- No more silent first-match: ambiguous selectors are rejected
Error model (new target-errors.ts):
- stale_ref: element identity changed since snapshot
- ambiguous: CSS selector matched multiple elements (with candidates)
- not_found: element not in DOM or invalid input
- All errors include actionable hints for AI agents
base-page.ts:
- click() and typeText() now use two-phase resolve-then-act
- Existing CDP fallback for click preserved
* feat: migrate scrollTo to unified resolver pipeline
scrollTo now uses the same two-phase resolve-then-act pattern as
click and typeText, getting fingerprint verification and structured
error diagnostics (stale_ref/ambiguous/not_found) for free.
* fix: address review — stronger fingerprint verification & surface TargetError in CLI
1. Fingerprint verification now uses the full identity vector (tag, id,
testId, ariaLabel, role, text) instead of just tag/role/text. Strong
identifiers (id, testId) are decisive; remaining signals use majority
voting. Fixes false negatives where same-tag elements swapped.
2. browserAction() now renders TargetError with code, hint, and
candidates list instead of just the message string.
* fix: migrate get/select/type-autocomplete to unified resolver
- browser get text/value/attributes now resolve via resolveTargetJs
instead of raw querySelector, getting fingerprint verification and
structured errors for free
- browser select uses selectResolvedJs on __resolved element
- type command's autocomplete detection uses isAutocompleteResolvedJs
on the already-resolved element
- Fix empty-string text prefix match: fp.text="Login" + text="" no
longer falsely passes fingerprint check
- docs/index.md: update feature cards to match README Highlights
- docs/zh/index.md: sync Chinese feature cards
- docs/guide/getting-started.md: align Highlights section
- README.zh-CN.md: rename "为什么是 OpenCLI" to "亮点", align with EN
* fix: remove duplicate extension zip from releases
The release and build-extension workflows were creating both
opencli-extension.zip and opencli-extension-v{version}.zip (identical
content), causing both to be uploaded. Keep only the versioned filename.
* docs: update extension zip filename to versioned format
Update all references from opencli-extension.zip to
opencli-extension-v{version}.zip to match the workflow change.
* feat(mubu): add mubu (mubu.com) adapter with 5 commands
Commands: doc, docs, notes, recent, search.
- Uses COOKIE strategy; API calls via in-page XHR with Jwt-Token
from localStorage (matches the web app's own mechanism).
- Renders node trees to Markdown (default) or plain text;
supports tables, tasks, images, emoji, mentions, strikethrough,
underline, and nested structures.
- notes supports flexible time ranges: single day, month, year,
or custom --from/--to spans, plus a --list overview mode.
- search returns full-text matches with hit count and snippets
for both folders and documents.
* fix(manifest): register mubu commands in runtime manifest
---------
Co-authored-by: jackwener <jakevingoo@gmail.com>
* refactor: unify OPENCLI_VERBOSE and DEBUG=opencli into one mechanism
Three debug output levels (verbose/debug/diagnostic) was redundant.
Merge DEBUG=opencli into OPENCLI_VERBOSE so `-v` flag controls all
verbose/debug output through a single mechanism.
- log.verbose() now checks both OPENCLI_VERBOSE and DEBUG=opencli
- log.debug() becomes an alias for log.verbose() (backward compat)
- boss/utils.js verbose helper simplified to check OPENCLI_VERBOSE only
- DEBUG=opencli still works as fallback (no breaking change)
* fix(boss): preserve debug fallback for verbose logs
The error path in executeCommand did not call page.closeWindow(),
leaving the automation window open until the extension's idle timer
fires. On Windows, MV3 service worker suspension makes this timer
unreliable, causing windows to linger indefinitely.
Now closeWindow is called after diagnostic collection but before
rethrowing, ensuring the window is closed on both success and failure.
Add missing env vars to both README and README.zh-CN:
- OPENCLI_SKIP_FETCH: skip adapter sync on global install
- OUTPUT: override output format (json/yaml/table)
- DEBUG=opencli: internal debug logging
- DEBUG_SNAPSHOT: DOM snapshot debug output
* fix: clean up stale .yaml adapter files from older versions (#953)
Users upgrading from v1.6.x retain .yaml adapter files in
~/.opencli/clis/ that trigger "Ignoring YAML adapter" warnings on
every run. The hash-based sync only tracks .js files, so these
legacy .yaml files are never cleaned up.
Add a cleanup step (3b) that removes .yaml/.yml files from user
adapter directories when the corresponding site exists in the
official package (i.e., the site has been migrated to .js).
* fix(fetch-adapters): narrow stale yaml cleanup
* feat: decouple extension version from CLI version
Extension and CLI had tightly coupled version numbers (both 1.7.2),
requiring manual sync across 3 files on every release. This decouples
them so each can release independently.
Changes:
- Extension version reset to 1.0.0 with independent versioning
- Extension sends compatRange (e.g. ">=1.7.0") in hello message
so doctor can check CLI/extension compatibility
- Daemon stores and exposes extensionCompatRange via /status
- Doctor uses compatRange for compatibility checks (falls back to
major-version check for older extensions without compatRange)
- Doctor shows extension update availability from cached GitHub
Releases data
- release.yml always builds and attaches extension zip to every
CLI release, so users always find both in the same release page
- build-extension.yml triggers on ext-v* tags (not v*) to avoid
duplicate builds
* fix: version extension release assets
* fix: code audit round 2 — pruneEmptyDirs, evaluateWithArgs, hot-reload, error cause chain
1. pruneEmptyDirs: use path.relative() instead of startsWith() to prevent
false boundary matches on overlapping directory names
2. evaluateWithArgs: add safe evaluate method that auto-serializes args via
JSON.stringify, preventing injection by design
3. Hot-reload: detect mtime changes on user adapter files in daemon mode,
invalidate module cache so edits take effect without restart
4. toEnvelope: preserve error cause chain in verbose mode for better
production debugging
* fix: address review feedback on code audit round 2
- pruneEmptyDirs: resolve() paths before relative() check
- evaluateWithArgs: validate keys are valid JS identifiers
- hot-reload: only bust ESM cache on reload, not first load
- toEnvelope: move cause serialization into toEnvelope itself
so all consumers (AI agents, MCP tools) get cause chain
* fix: address code audit findings (C1-C4, I1, I4, I6)
Security:
- C1: Fix page.evaluate injection in browser type/select commands and
6 adapter files by using JSON.stringify for user input interpolation
- C2: Close WebSocket on CDP connect timeout to prevent resource leak
- C3: Reject CDP connect promise on Page.enable failure instead of
silently swallowing the error
Reliability:
- C4: Guard against corrupted adapter-manifest.json hashes to prevent
false-positive override deletion
- I1: Throw on pre-navigation failure instead of warn-and-continue
- I4: Use Map<string, Promise<void>> for lazy module loading to prevent
concurrent double-imports of the same adapter
Performance:
- I6: Replace O(n) registry alias cleanup with O(k) direct deletion
* fix: address self-review findings on PR #981
- C1: add quotes around CSS selector attribute values in browser
type/select to match other commands (get text/value/attributes)
- C2: clear this._ws in timeout handler to prevent race with open event
- C4: refine corruption guard — treat null/undefined hashes as empty,
only skip sync for truly invalid types (string, number, array)
* feat(clis/chatgptweb): add ChatGPT web image generation command
Add `opencli chatgptweb image` command that generates images using
ChatGPT web (GPT-4o image generation) and saves them locally.
Features:
- Navigates to chatgpt.com/new with full page reload to ensure clean state
- Uses Playwright's page.type() for reliable text input in TipTap editor
- Closes sidebar if open (covers the chat composer on some layouts)
- Polls for response completion (handles thinking/throttling states)
- Extracts generated images from DOM (backend-api/estuary/content URLs)
- Downloads and saves as PNG/JPEG files to user-specified directory
- Supports --op for output directory and --sd to skip download
Files:
- clis/chatgptweb/image.js: CLI command definition
- clis/chatgptweb/utils.js: DOM helpers, send/wait/export functions
Works cross-platform (Linux/macOS/Windows) via OpenCLI browser automation.
* fix(chatgptweb): stabilize image generation flow
* docs(chatgptweb): add browser adapter guide
---------
Co-authored-by: Tony Simons <tony@tonysimons.dev>
Co-authored-by: jackwener <jakevingoo@gmail.com>
Support browsing second-hand houses, neighborhoods, rentals, and
transaction records on ke.com with city/district/price filtering.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* feat(maimai): add talent search with multi-dimensional filters
Add maimai.cn talent search adapter with support for:
- Keyword search (query)
- Company filtering (multiple companies supported)
- School filtering (with 985/211 options)
- Location filtering (province/city)
- Work experience and education level filters
- Industry and position filters
- Direct chat availability
- Sort by relevance, activity, work years, or education
Features:
- Reuses Chrome login session for authentication
- Extracts candidate info: name, job title, company, work history
- Shows work years, education, age, active status
- Displays skill tags and mutual friends count
* fix docs and strategy for maimai adapter
---------
Co-authored-by: jackwener <jakevingoo@gmail.com>
* feat(discord-app): add delete command to remove a message by ID
Adds a new `delete` command for the discord-app CLI that deletes a
message in the active channel by its snowflake ID. Uses the UI strategy
to hover the message, open the "More" menu, click "Delete Message", and
confirm the deletion dialog.
* docs: add binance adapter doc and update discord doc with delete command
* feat(twitter): add lists command to retrieve user lists
Add twitter/lists command that fetches Twitter/X lists for a user.
Supports:
- Lists with member and follower counts
- Private/public mode detection
- Default to current user if no user specified
- Works for any Twitter user
* docs: add lists command to twitter commands in README
Add twitter lists command to Built-in Commands table in both
English and Chinese README files
* fix(twitter): parse lists from card DOM instead of locale-specific page text
---------
Co-authored-by: isanwenyu <isanwenyu@users.noreply.github.com>
Co-authored-by: jackwener <jakevingoo@gmail.com>
* fix(zsxq): accept topic_id as string in getTopicFromResponse
The ZSXQ API returns topic_id as a string, but getTopicFromResponse()
only checked for typeof === 'number', causing it to fall through and
return null. This made 'opencli zsxq topic <id>' fail with NOT_FOUND
for all valid topic IDs.
* fix(zsxq): use group-scoped topic endpoint instead of bare /v2/topics/{id}
The ZSXQ API requires topics to be fetched within their group context.
Change /v2/topics/{id} -> /v2/groups/{groupId}/topics/{id} for both
the detail and comments endpoints. Also adds optional --group_id arg.
* fix: include adapter tests in default npm test
`npm test` only ran unit + extension projects, so adapter tests
(clis/**/*.test.js) were never exercised by the default test command.
Add --project adapter so they run alongside unit and extension tests.
* test: include adapter project in default npm test
* refactor: smart sync adapters instead of full copy (#sparse-override)
Replace unconditional full-copy of all adapters to ~/.opencli/clis/ with
hash-based smart sync that only copies files whose content has changed.
Changes:
- fetch-adapters.js: use SHA-256 content hashes to skip unchanged files;
store per-file hashes in adapter-manifest.json
- discovery.ts: simplify ensureUserAdapters() to only create the directory
(no longer triggers full copy on first run)
- main.ts: fix fast completion to check manifest file existence instead of
directory existence (sparse override may have empty user dir)
- cli.ts: add `opencli adapter eject/reset/status` commands for managing
local adapter overrides
- engine.test.ts: add tests for empty user dir and ensureUserAdapters
* fix: address review blockers — site-level sync + reset --all
1. Fix `adapter reset --all`: change <site> from required to optional
argument so --all can be used without specifying a site name.
2. Change smart sync from file-level to site-level granularity:
if any file in a site has changed upstream, overwrite the entire
site directory. This matches the agreed product semantics — local
modifications to any file in a site are replaced when upstream
updates that site.
* fix: delete old site dir before writing updated adapter files
When a site has upstream changes, delete the entire site directory
first, then write the new version. This prevents stale files from
older versions lingering in the user directory.
* fix: reset --all preserves custom sites, only removes official overrides
Blocker 3 fix: reset --all now checks BUILTIN_CLIS to identify official
sites and only deletes those, preserving user-created custom sites.
* refactor: sparse sync deletes local overrides instead of copying new versions
Changed fetch-adapters.js semantics per team agreement:
- When an official site has upstream changes, DELETE the local override
instead of copying the new version into ~/.opencli/clis/
- Runtime automatically falls back to package baseline
- ~/.opencli/clis/ becomes a true sparse override layer
* fix: reset <site> rejects custom sites, only allows official overrides
Single-site reset now checks BUILTIN_CLIS before deleting, matching
the same protection that reset --all already has.
* fix: reset <site> allows custom sites per product decision
Per @WAWQAQ: explicit single-site reset should work on custom sites too.
Differentiate messaging: official sites say "using official baseline",
custom sites say "removed custom site".
reset --all still only removes official overrides (bulk safety).
* fix: reset --all deletes all local sites including custom per product decision
Per @WAWQAQ: --all should clear the entire local working cache,
including custom sites. Single-site reset already handles both types.
Binance was the only adapter left in src/clis/ after the TS→JS
migration (PR #928). Move all 11 adapters and the test file to
clis/binance/, strip TypeScript syntax from the test, and switch
the test import to the @jackwener/opencli/pipeline package export.
log.debug() requires DEBUG=opencli to output, which means
DEBUG_SNAPSHOT=1 alone no longer shows snapshot fallback diagnostics.
Use process.stderr.write directly since the DEBUG_SNAPSHOT guard
already controls when this diagnostic fires.
Users who created custom .ts adapters in ~/.opencli/clis/ will see
their commands silently disappear after upgrading to the JS-only
version. Add an explicit warning so they know to convert to .js.
The alias resolution logic checked `!registry.has(target)` before
calling `registry.get(target)`, which always returned undefined.
Moreover, aliases registered as `site/alias` keys meant `registry.has`
returned true, skipping the block entirely. The canonical name was
never resolved, so `validate site/alias` silently checked 0 commands.
Simplify to always resolve via `registry.get(target)` which handles
both canonical keys and alias keys correctly.
Older versions (pre-1.7.1) shipped adapters as .ts files. When users
upgrade to a .js-only version, the old .ts files are left orphaned in
~/.opencli/clis/. Add a cleanup step that removes .ts files when a
corresponding .js official adapter exists.
XavierCai