Token导航 LogoToken导航TokenDH.com
研究检索敏感数据github未标认证来源可访问许可证需确认审计提醒

yandex-directyandex 直接

Agent Skill

yandex-direct 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

679

周安装

28

GitHub Stars

公开资料未说明

下载量

222
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:yandex-direct(yandex 直接)
来源仓库:https://github.com/elsvv/yandex-direct-skill
仓库路径:skills/yandex-direct
安装命令:
npx skills add https://github.com/elsvv/yandex-direct-skill --skill yandex-direct
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/elsvv/yandex-direct-skill --skill yandex-direct

简介

yandex-direct 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。

  • 它基于关键词或场景线索提供信息聚合与筛选能力,适用于研究类任务。
  • 通过 npx skills add 命令从 GitHub 仓库安装,具体实现细节请查阅原始文档。
  • 使用前建议确认权限范围、项目维护情况以及是否会触发联网或文件读写等操作。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Yandex Direct API v5

Essentials

Base URLs

EnvironmentJSON EndpointSOAP Endpoint
Productionhttps://api.direct.yandex.com/json/v5/{service}https://api.direct.yandex.com/v5/{service}
Sandboxhttps://api-sandbox.direct.yandex.com/json/v5/{service}https://api-sandbox.direct.yandex.com/v5/{service}
Reportshttps://api.direct.yandex.com/json/v5/reportsSame pattern
Reports (sandbox)https://api-sandbox.direct.yandex.com/json/v5/reportsSame pattern

Authentication

All requests require an OAuth token in the Authorization header.

Required Headers (every request):

HeaderValueRequired
AuthorizationBearer YOUR_OAUTH_TOKENAlways
Content-Typeapplication/json; charset=utf-8Always
Accept-Languageru or enRecommended
Client-Loginclient_loginAgency accounts only

Getting an OAuth Token:

  1. Register app at https://oauth.yandex.ru/client/new with direct:api permission
  2. Get token: https://oauth.yandex.ru/authorize?response_type=token&client_id=YOUR_CLIENT_ID
  3. Token appears in redirect URL: #access_token=TOKEN&token_type=bearer&expires_in=31536000
  4. Token is valid for 1 year

Quick token setup:

bash scripts/get_token.sh --client-id YOUR_CLIENT_ID

Request Format (JSON)

Every API request (except Reports) follows this JSON structure:

{
  "method": "get|add|update|delete|suspend|resume|archive|unarchive|...",
  "params": {
    "SelectionCriteria": { ... },
    "FieldNames": ["Id", "Name", ...],
    "Page": { "Limit": 10000, "Offset": 0 }
  }
}

Key patterns:

  • SelectionCriteria -- filter which objects to return (Ids, CampaignIds, States, Statuses, Types, etc.)
  • FieldNames -- which fields to include in the response
  • Page -- pagination: Limit (max 10000) and Offset
  • add methods use an array of objects (e.g., "Campaigns": [...])
  • update methods use an array of objects with Id field
  • delete methods use SelectionCriteria with Ids array

Response Format

{
  "result": {
    "Campaigns": [ ... ],
    "LimitedBy": 10000
  }
}
  • LimitedBy appears when there are more objects than returned (need pagination)

Error response:

{
  "error": {
    "error_code": 53,
    "error_string": "Authorization error",
    "error_detail": "Token not found or expired"
  }
}

Units (API Points) System

Every response includes the Units HTTP header: Units: spent/remaining/daily_limit

Example: Units: 10/20828/64000 means 10 points spent, 20828 remaining, 64000 daily limit.

RuleDetail
Daily limitIndividual per advertiser, based on campaign activity
RefreshPoints awarded every 60 minutes (sliding 24h window)
Per period1/24 of daily limit per hour + unspent from previous 23 hours
Concurrent requestsMax 5 simultaneous requests per advertiser
Error cost20 points per error (except server errors)
Minimum daily limit~64,000 points for active accounts
AgencyPoints deducted from advertiser by default; agency can opt to use own points

Common Error Codes

CodeMeaningAction
53Authorization errorCheck token validity
152Insufficient pointsWait for points refresh (hourly)
1000Concurrent request limitReduce parallel requests
1001Operation limit exceededReduce batch size
1002Invalid tokenReauthorize, get new token
2000Unknown errorRetry after delay
8800Object limit per requestReduce batch size
9000Insufficient unitsWait for daily limit refresh

Sandbox

  • Completely isolated from production data
  • No web interface -- API only
  • Same restrictions as live API
  • Reports limited to one campaign per request
  • Data deleted after 1 month of inactivity
  • Roles: Advertiser or Agency (with 3 test clients)
  • Use YANDEX_DIRECT_SANDBOX=true in config or set base URL to api-sandbox.direct.yandex.com

Configuration

The skill uses config/.env for credentials:

# Required
YANDEX_DIRECT_TOKEN=your_oauth_token

# Optional: sandbox mode
YANDEX_DIRECT_SANDBOX=true

# Optional: agency client login
YANDEX_DIRECT_CLIENT_LOGIN=client_login

Scripts

IMPORTANT: Always run scripts with bash prefix and absolute paths from the skill directory. Scripts use bash-specific features and will not work if sourced from zsh. Do NOT source scripts/common.sh directly — use the wrapper scripts below.

check_connection.sh

Verify API token and list available campaigns.

bash scripts/check_connection.sh

campaigns.sh

Manage campaigns: list, get details, stats, suspend/resume/archive.

# List all campaigns
bash scripts/campaigns.sh --action list

# List only active campaigns
bash scripts/campaigns.sh --action list --states ON

# Get full campaign details
bash scripts/campaigns.sh --action get --ids 12345678

# Get campaign funds/spend info
bash scripts/campaigns.sh --action stats --ids 12345678

# Suspend a campaign
bash scripts/campaigns.sh --action suspend --ids 12345678

# Resume a campaign
bash scripts/campaigns.sh --action resume --ids 12345678
ParamDescription
--action, -alist, get, stats, suspend, resume, archive, unarchive
--ids, -iComma-separated campaign IDs
--states, -sFilter: ON, OFF, SUSPENDED, ENDED, ARCHIVED
--limit, -lMax results (default: 100)

ads.sh

Manage ads: list, get details, suspend/resume/moderate/archive.

# List ads by campaign
bash scripts/ads.sh --action list --campaign-ids 12345678

# List ads by ad group
bash scripts/ads.sh --action list --adgroup-ids 987654

# Get specific ad details
bash scripts/ads.sh --action get --ad-ids 111222333

# Suspend an ad
bash scripts/ads.sh --action suspend --ad-ids 111222333
ParamDescription
--action, -alist, get, suspend, resume, moderate, archive
--campaign-idsFilter by campaign IDs
--adgroup-idsFilter by ad group IDs
--ad-idsSpecific ad IDs
--limit, -lMax results (default: 1000)

keywords.sh

Manage keywords and autotargeting.

# List keywords by campaign
bash scripts/keywords.sh --action list --campaign-ids 12345678

# List keywords by ad group
bash scripts/keywords.sh --action list --adgroup-ids 987654

# Suspend keywords
bash scripts/keywords.sh --action suspend --ids 111,222,333

reports.sh

Pull statistics reports (campaign, ad group, ad, keyword level).

# Campaign performance (last 30 days)
bash scripts/reports.sh

# Campaign performance for custom date range
bash scripts/reports.sh --date-range CUSTOM_DATE --date-from 2026-01-01 --date-to 2026-01-31

# Ad-level performance
bash scripts/reports.sh --type AD_PERFORMANCE_REPORT \
  --fields "AdId,AdGroupId,Impressions,Clicks,Ctr,AvgCpc,Cost"

# Keyword/criteria performance with filter
bash scripts/reports.sh --type CRITERIA_PERFORMANCE_REPORT \
  --fields "CriteriaType,Criteria,Impressions,Clicks,Ctr,AvgCpc,Cost" \
  --filter '{"Field":"CampaignId","Operator":"EQUALS","Values":["12345678"]}'

# Demographics (age/gender)
bash scripts/reports.sh --type CAMPAIGN_PERFORMANCE_REPORT \
  --fields "Age,Gender,Impressions,Clicks,Ctr,Cost"

# Search queries report
bash scripts/reports.sh --type SEARCH_QUERY_PERFORMANCE_REPORT \
  --fields "Query,Impressions,Clicks,Ctr,Cost"

# Save report to file
bash scripts/reports.sh --output report.tsv

# Predefined date ranges
bash scripts/reports.sh --date-range YESTERDAY
bash scripts/reports.sh --date-range LAST_7_DAYS
bash scripts/reports.sh --date-range THIS_MONTH
ParamDefaultDescription
--type, -tCAMPAIGN_PERFORMANCE_REPORTReport type (see below)
--date-range, -rLAST_30_DAYSDate range preset
--date-fromStart date for CUSTOM_DATE
--date-toEnd date for CUSTOM_DATE
--fields, -fCampaignName,Impressions,Clicks,Ctr,AvgCpc,CostFields
--filterFilter JSON
--output, -ostdoutSave to file
--name, -nauto-generatedReport name

Report types: ACCOUNT_PERFORMANCE_REPORT, CAMPAIGN_PERFORMANCE_REPORT, ADGROUP_PERFORMANCE_REPORT, AD_PERFORMANCE_REPORT, CRITERIA_PERFORMANCE_REPORT, SEARCH_QUERY_PERFORMANCE_REPORT, CUSTOM_REPORT

Date ranges: TODAY, YESTERDAY, LAST_3_DAYS, LAST_7_DAYS, LAST_14_DAYS, LAST_30_DAYS, LAST_90_DAYS, THIS_MONTH, LAST_MONTH, ALL_TIME, CUSTOM_DATE

dictionaries.sh

Get reference data (regions, currencies, etc.).

bash scripts/dictionaries.sh --dict GeoRegions
bash scripts/dictionaries.sh --dict Currencies

Advanced: common.sh functions

For custom API calls not covered by scripts above, use common.sh functions. Must be run inside a bash script (not sourced from zsh):

#!/bin/bash
source /path/to/scripts/common.sh
load_config
response=$(direct_request "campaigns" '{"method":"get","params":{...}}')

All API v5 Services

ServiceEndpoint SuffixMethodsPurpose
Campaigns/campaignsadd, update, delete, get, suspend, resume, archive, unarchiveCampaign management
AdGroups/adgroupsadd, update, delete, getAd group management
Ads/adsadd, update, delete, get, moderate, suspend, resume, archive, unarchiveAd management
Keywords/keywordsadd, update, delete, get, suspend, resumeKeyword/autotargeting management
BidModifiers/bidmodifiersadd, set, delete, getBid adjustment management
KeywordBids/keywordbidsset, setAuto, getKeyword bid management
AudienceTargets/audiencetargetsadd, delete, suspend, resume, get, setBidsAudience target management
RetargetingLists/retargetinglistsadd, update, delete, getRetargeting list management
Sitelinks/sitelinksadd, delete, getSitelink set management
AdExtensions/adextensionsadd, delete, getCallout extension management
VCards/vcardsadd, delete, getVirtual business card management
AdImages/adimagesadd, delete, getImage management
AdVideos/advideosadd, getVideo management
Creatives/creativesadd, getCreative management
Reports/reportsPOST (custom)Statistics and reporting
Dictionaries/dictionariesgetReference data (regions, currencies, etc.)
Clients/clientsget, updateAdvertiser account management
AgencyClients/agencyclientsadd, update, getAgency client management
Changes/changescheck, checkCampaigns, checkDictionariesChange tracking
Feeds/feedsadd, update, delete, getProduct feed management
DynamicTextAdTargets/dynamictextadtargetsadd, delete, get, resume, setBids, suspendDynamic ad targeting
SmartAdTargets/smartadtargetsadd, update, delete, get, resume, setBids, suspendSmart banner targeting
TurboPages/turbopagesgetTurbo page parameters
Businesses/businessesgetBusiness profile data
Strategies/strategiesadd, update, get, archive, unarchivePortfolio strategy management
NegativeKeywordSharedSets/negativekeywordsharedsetsadd, update, delete, getShared negative keyword sets
KeywordsResearch/keywordsresearchhasSearchVolume, deduplicateKeyword preprocessing
Leads/leadsgetTurbo page form submissions
Bids/bidsset, setAuto, getBid management (legacy)

Detailed References

Read the reference file matching the area you need:

Guidelines

  • Always verify the token is valid before batch operations: bash scripts/check_connection.sh
  • Use Page.Limit and Page.Offset for paginating through large result sets (max 10000 per request)
  • Request only the FieldNames you actually need to save API points
  • Use the Changes service to detect modifications before re-downloading all data
  • For reports, use processingMode: auto and handle both 200 (ready) and 201/202 (pending) HTTP status codes
  • Set returnMoneyInMicros: false in report headers to get human-readable monetary values
  • For agency accounts, always include the Client-Login header
  • Max 5 concurrent requests per advertiser -- queue or throttle your requests
  • Use sandbox (api-sandbox.direct.yandex.com) for development and testing
  • Store tokens securely in config/.env (file is gitignored)
  • For monetary values in standard API requests/responses: amounts are in micros (multiplied by 1,000,000)
  • For monetary values in reports: use returnMoneyInMicros: false header to get human-readable values, or divide by 1,000,000

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.53%
按下载量换算79

Claude

29.15%
按下载量换算65

Cursor

19.71%
按下载量换算44

Gemini CLI

9.73%
按下载量换算22

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills