x-agent

TotalClaw 作者 totalclaw

以三种模式规划和运行 X (Twitter) 操作：(1) 仅监控智能，(2) 起草并批准发布，以及 (3) 具有严格护栏的有限自动化。在设置 X 帐户工作流程、起草帖子/回复/话题、定义安全策略、安排内容或审查绩效时使用。

安装 / 下载方式

TotalClaw CLI推荐

totalclaw skills install shop:totalclaw:totalclaw~inceptivco-x-agent

cURL直接下载，无需登录

curl -fsSL https://skills.taituai.com/api/skills/totalclaw%3Atotalclaw~inceptivco-x-agent/file -o inceptivco-x-agent.md

## 概述（中文）

## 原文

# X Agent (Phased Setup)

## Objective
Operate an X workflow safely across three phases, with automation disabled until explicitly enabled.

## Phase 1 — Monitor-only
1. Track topics, keywords, and accounts.
2. Summarize notable posts and sentiment.
3. Produce a daily/weekly brief:
- top signals
- opportunities
- reputational risks
4. Do not post.

## Phase 2 — Draft + approval
1. Generate draft tweets/replies/threads from user goals.
2. Enforce style rules:
- concise
- no unverifiable claims
- no legal/financial guarantees
3. Present final draft with:
- risk flags
- confidence level
- suggested posting window
4. Require explicit approval before publishing.

## Phase 3 — Limited automation (default OFF)
Enable only after user confirmation.

Automation constraints:
1. Allowed actions only from pre-approved playbooks.
2. Hard caps:
- max posts/day
- max replies/hour
3. Blocked categories:
- politics (unless explicitly allowed)
- legal/medical/financial advice language
4. Quiet hours and cooldowns required.
5. Global kill switch required.

## Required Guardrails
Always define these before posting access:
1. Account scope: monitor | draft+approve | limited-auto
2. Max daily posts
3. Max hourly replies
4. Quiet hours (timezone)
5. Banned topics/phrases
6. Kill switch command and owner

## Starter Playbooks
1. News reaction post (single tweet)
2. Research summary thread (3–5 posts)
3. Reply triage (approve queue)
4. Daily recap post

## Rollout Checklist
1. Create or choose X account.
2. Configure API credentials in local environment (never in chat).
3. Start in monitor-only for 3–7 days.
4. Move to draft+approve for 1–2 weeks.
5. Enable limited automation for one playbook only.
6. Review weekly and adjust limits.

## Output Formats
### Monitor brief
- What happened
- Why it matters
- Suggested response

### Draft package
- Draft text
- Variant A/B
- Risk flags
- Recommended time

### Automation report
- Actions executed
- Actions skipped (and why)
- Rate limit usage
- Any policy violations blocked