Alibaba Cloud cleversee CLI skill. web-search is a general-purpose, AI/agent-oriented
internet search providing high-quality, low-latency neural knowledge retrieval and web
content scraping across any topic and real-time information.
Triggers: "cleversee", "web search", "AI 搜索", "联网搜索", "查资料", "实时信息".
cleversee is a standalone CLI tool for Alibaba Cloud users that provides:
Web Search (web-search): AI-powered internet search for agents — high-quality, low-latency retrieval of relevant results and page content across any topic and real-time information, with optional date-range and domain filtering. It creates no cloud resources and requires no cleanup.
Interactive commands (cannot be used in non-interactive/automated contexts):
cleversee auth set (no parameters)
cleversee auth switch
cleversee auth --mode OAuth — OAuth login; skips the mode-selection prompt but still opens a browser, so it requires a present user
RAM Policy
This skill requires AliyunCleverSeeAISearchPlatformUserAccess permission.
See references/ram-policies.md for the full permission list and policy document.
[MUST] Permission Failure Handling: When any command or API call fails due to permission errors at any point during execution, follow this process:
Read references/ram-policies.md to get the full list of permissions required by this SKILL
Use ram-permission-diagnose skill to guide the user through requesting the necessary permissions
Pause and wait until the user confirms that the required permissions have been granted
Parameter Confirmation
IMPORTANT: Parameter Confirmation — Before executing any command or API call,
ALL user-customizable parameters (e.g., search query, limit, search-type, region,
date ranges, domain lists, profile name, etc.) MUST be confirmed with the
user. Do NOT assume or use default values without explicit user approval.
Web Search Parameters
Parameter
Required
Default
Description
--query <query>
Yes
—
Search keywords
--limit <n>
No
10
Result count, range 1-10
--search-type <type>
No
pro
pro (advanced, global sources) or lite (fast, low-cost)
--region <region>
No
mainland_china
mainland_china or global (pro mode only).
--start-time <date>
No
—
Start date YYYY-MM-DD
--end-time <date>
No
—
End date YYYY-MM-DD
--include-domain <d1 d2...>
No
—
Domain whitelist, space-separated (pro only; forces --region global automatically)
--exclude-domain <d1 d2...>
No
—
Domain blacklist, space-separated (pro only)
--table
No
—
Output as a table view (default web-search view is a card list)
Global Parameters
Apply to web-search (and --profile also applies to auth). doctor accepts no options.
Parameter
Default
Description
--profile <name>
defaultProfile
Aliyun CLI profile name
-o, --output <json|table>
table
Output format (table for human-readable, json for program parsing)
-V, --version
—
Print cleversee version
Core Workflow
Workflow A: Web Search
text
Step 1: Verify environment → cleversee doctor
Step 2: Confirm search params → Ask user for query, limit, search-type, region, dates, domains
Step 3: Execute search → cleversee web-search --query "<user-confirmed query>" [options] -o json
Step 4: Parse results → Check JSON output for at least one result
Example (parameters must be confirmed with user first):
bash
# Step 1: Verify environment
cleversee doctor
# Step 2-3: Execute search with user-confirmed parameters
cleversee web-search \
--query "What are the latest battery breakthroughs in electric vehicles?" \
--limit 5 \
--search-type pro \
-o json
Workflow B: Credential Management
text
Step 1: Check current status → cleversee doctor
Step 2: Configure credentials → cleversee auth set --mode <mode> [params]
Step 3: Verify → cleversee doctor
Example:
bash
# Step 1: Check current status
cleversee doctor
# Step 2: Configure RamRoleArn (parameters provided by user)
cleversee auth set --mode RamRoleArn \
--access-key-id <user-provided-id> \
--access-key-secret <user-provided-secret> \
--arn <user-provided-arn>
# Step 3: Verify
cleversee doctor
Success Verification
Quick checks:
cleversee doctor exits with code 0 and all checks pass
cleversee web-search exits with code 0 and returns at least one result
Cleanup
cleversee web-search is a read-only operation — it creates no cloud resources and requires no cleanup.
For credential cleanup, reconfigure or remove profiles via cleversee auth set or Aliyun CLI profile management.
Best Practices
Run cleversee doctor first — always verify environment health before searching
Use -o json — for programmatic result parsing
Run cleversee doctor — to diagnose environment or credential issues when a command fails
Use --search-type lite — for quick searches that don't need advanced filtering
Use --include-domain / --exclude-domain — in pro mode to focus on authoritative sources
Use --start-time / --end-time — to get recent content within a specific date range
Use cleversee auth set for automation — never use interactive auth or auth switch in non-interactive contexts
Use --profile — manage multiple account credentials
Use --help — every command level supports --help for accurate parameter discovery
