运营数据日报虾

运营数据日报虾 — 自动化多平台运营数据采集与日报生成。从抖音、小红书、视频号、B站、微博等平台采集数据，清洗标准化后生成结构化日报，并推送到飞书文档/群聊。 **当以下情况时使用此 Skill**： (1) 需要汇总多个内容平台（抖音/小红书/视频号/B站/微博）的运营数据 (2) 需要生成运营日报、周报、月报...

Security Scan

The skill's purpose (collecting platform metrics and generating reports) is plausible, but the package metadata omits the many sensitive credentials the included scripts actually require and there are small scope mismatches — proceed only after addressing those discrepancies and reviewing secrets handling.

Purpose & Capability

The skill claims to aggregate platform metrics and push reports to Feishu, which matches the included scripts. However, the registry metadata lists no required environment variables or primary credential, while the scripts and reference docs clearly expect multiple sensitive credentials (DOUYIN_ACCESS_TOKEN, XIAOHONGSHU_COOKIE, WEIXIN_CORP_ID/SECRET, BILIBILI_SESSDATA, WEIBO_ACCESS_TOKEN, etc.). That omission is an inconsistency: a legitimate aggregator should declare these requirements up front.

Instruction Scope

SKILL.md and the scripts keep to the stated workflow: ask user which platforms, read credentials from a local .env, call platform APIs (curl), normalize JSON, run report generator, then push via Feishu tools. The instructions reference reading a .env and saving files under data/raw and data/reports (expected). They also instruct use of feishu_create_doc and message tools (agent/tool integration) and scheduling via cron. Nothing in the instructions reads unrelated system config or exfiltrates to unknown endpoints.

Install Mechanism

There is no install spec (lower install risk) but the skill bundle includes executable scripts that will run on the host. The fetch script uses curl/jq/python and the report generator requires pandas/jinja2; the package does not declare dependency installation. This is not malicious but means the operator must ensure required binaries and Python packages are installed in a controlled environment.

Credentials

The skill requires multiple high-sensitivity credentials (OAuth tokens, cookies, SESSDATA, corp secrets) for the platforms it integrates with. The registry metadata declares no required env vars — a mismatch that hides the true secrets surface. Using cookies (e.g., XIAOHONGSHU_COOKIE or BILIBILI_SESSDATA) is inherently fragile and sensitive and should be minimized or rotated. The number and sensitivity of env vars is proportionate to the task, but they should be declared and justified in metadata.

Persistence & Privilege

The skill is user-invocable and not forced-always; it does not request elevated platform privileges or modify other skills. It reads/writes files under its own workspace (data/) and suggests optional cron scheduling, which is normal for automation. Autonomous invocation is allowed (platform default) but not combined with other red flags here.

What to consider before installing

This skill appears to implement the advertised report workflow, but the package metadata hides the real secrets it needs. Before installing: (1) insist the publisher update metadata to list required env vars (DOUYIN_*, XIAOHONGSHU_COOKIE, WEIXIN_CORP_*, BILIBILI_SESSDATA, WEIBO_*, etc.); (2) review scripts locally (they call only known platform endpoints via curl) and run them in an isolated environment or container; (3) avoid pasting long-lived full-account credentials—prefer scoped tokens where possible and rotate cookies/tokens frequently; (4) confirm how Feishu integration is performed (the skill expects external agent tools like feishu_create_doc/message); (5) if you must run on production servers, ensure data/ and .env are permission-restricted and consider limiting network egress or running behind a proxy to monitor outgoing API calls. If the publisher cannot justify the missing metadata or answer how credentials are stored/rotated, treat the skill as untrusted.

自动化多平台运营数据采集 → 清洗分析 → 日报生成 → 飞书推送。

工作流程

步骤 1：确认配置

询问用户：

• 需要采集哪些平台（抖音/小红书/视频号/B站/微博）
• 账号 ID 和 API 密钥是否已配置（参考 references/platform-api-guide.md）
• 日报类型：日报 / 周报 / 月报
• 推送目标：飞书群 chat_id 或个人

步骤 2：数据采集

使用 scripts/fetch-platform-data.sh 采集数据：

# 采集单平台
./scripts/fetch-platform-data.sh fetch douyin <account_id>

# 采集所有平台（并行）
./scripts/fetch-platform-data.sh fetch all

数据输出为 JSON 格式，存储在 data/raw/<platform>_<date>.json。

步骤 3：生成日报

使用 scripts/generate-report.py 生成报告：

# 生成昨日日报
python3 scripts/generate-report.py --date yesterday

# 生成指定日期
python3 scripts/generate-report.py --date 2026-03-31

# 生成周报
python3 scripts/generate-report.py --type weekly --date 2026-03-31

步骤 4：推送到飞书

1. 用 feishu_create_doc 创建飞书云文档（Markdown 格式）
2. 用 message 工具将文档链接发送到目标群聊或个人
3. 若有异常数据，额外发送 @全体成员 的预警消息

步骤 5：配置定时任务（可选）

若用户需要每天自动执行，提供 cron 配置示例：

# 每天早上 8:00 自动生成日报
0 8 * * * cd /path/to/workspace && ./scripts/fetch-platform-data.sh fetch all && python3 scripts/generate-report.py --date yesterday

数据标准字段

所有平台数据统一标准化为：

字段	说明
platform	平台名称
date	数据日期
views	播放量/阅读量
likes	点赞数
comments	评论数
shares	转发/分享数
favorites	收藏数
new_followers	当日涨粉数
total_followers	粉丝总数
engagement_rate	互动率（(likes+comments+shares)/views）
completion_rate	完播率（视频平台）

异常检测

采集完成后自动检查异常，规则详见 references/anomaly-rules.md。 发现异常时在日报中标注 ⚠️，并单独发送预警消息。

日报模板

日报格式详见 references/report-templates.md，支持：

• 简洁版（仅核心指标）
• 详细版（含趋势图）
• 对比版（多平台横向对比）
• 异常版（仅展示异常数据）

平台 API 接入

各平台 API 鉴权和调用方式详见 references/platform-api-guide.md。

注意事项

• 小红书无官方 API，需 Cookie 鉴权，稳定性较低，建议每周检查 Cookie 是否过期
• API 调用有频率限制，建议每天采集 1-2 次
• 确保服务器时区与平台数据时区一致（Asia/Shanghai）
• 历史数据存储在本地 SQLite，长期使用需定期备份