Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add anneheartrecord/charles-skill --skill x-content-reviewgit clone --depth 1 https://github.com/anneheartrecord/charles-skillWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/anneheartrecord/charles-skill/x-content-review)<a href="https://agentmods.dev/skills/anneheartrecord/charles-skill/x-content-review"><img src="https://agentmods.dev/badge/skills/anneheartrecord/charles-skill/x-content-review/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/anneheartrecord/charles-skill/x-content-review"><img src="https://agentmods.dev/badge/skills/anneheartrecord/charles-skill/x-content-review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00171 | $0.02439 |
| Opus 5 | $0.00086 | $0.01220 |
| Sonnet 5 | $0.00034 | $0.00488 |
| Haiku 4.5 | $0.00017 | $0.00244 |
Grade A, and why
x-content-review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
X 数据内容复盘
把 X Analytics 数据变成可执行的运营决策:什么类别涨粉快、什么时段发效率高、哪些帖子该扩成系列、哪些该换头重发、哪些选题该停。
数据源(三条路,按优先级)
- API 模式(首选,配了官方 API 凭证时):X 官方 pay-per-use 接口读自己账号数据,$0.001/请求,一次复盘不到 1 美分。官方接口稳定,且
impression_count是真曝光数,口径与 analytics 后台对齐。凭证为四个 env(X_API_KEY/X_API_SECRET/X_ACCESS_TOKEN/X_ACCESS_TOKEN_SECRET),依赖 tweepy。 - Pulse 模式(免费兜底):cookie 登录态(env
X_AUTH_TOKEN/X_CT0,依赖 twscrape)拉公开指标,views 为公开浏览数。 - CSV 模式(增强,用户主动导出时):X Analytics 后台两份 CSV(
account_overview_analytics*.csv+account_analytics_content*.csv),独有「净涨粉/主页访问」,转化口径升级为「涨粉/万曝光」。运营目录里有 7 天内的新 CSV 对时自动改走 CSV 模式。
API 和 pulse 模式的涨粉趋势都靠粉丝数快照 JSONL 差值(每次跑自动追加快照,越用越准),转化口径为「收藏/万曝光」。
工作流程
第 1 步:选数据路径
- 先看用户运营目录(惯例
30-outputs/运营/)有无 7 天内导出的 CSV 对:有 → CSV 模式; - 没有,四个 API env 齐 → API 模式:
python3 <skill目录>/scripts/fetch_x_api.py --limit 100 \
--snapshot-file <运营目录>/x-follower-snapshots.jsonl > /tmp/x-pulse.json
- API 凭证不全 → pulse 模式:
python3 <skill目录>/scripts/fetch_x_pulse.py --user <handle> --limit 100 \
--snapshot-file <运营目录>/x-follower-snapshots.jsonl > /tmp/x-pulse.json
两个脚本输出同构 JSON,后续分析命令一样。凭证缺失时脚本会报错并给配置指引。
第 2 步:跑分析脚本
# API / pulse 模式(两者输出同构,统一走 --pulse)
python3 <skill目录>/scripts/analyze_x_data.py --pulse /tmp/x-pulse.json \
--snapshots <运营目录>/x-follower-snapshots.jsonl --days 7
# CSV 模式
python3 <skill目录>/scripts/analyze_x_data.py \
--overview <最新overview.csv> --content <最新content.csv> --days 7
脚本输出 JSON(两种模式同构):
content.time_slots_beijing:北京时间 2 小时分桶的时段表现(CSV 模式从 Post id 雪花 ID 反推发帖时间)content.categories:类别转化,reply 单列;pulse 口径「收藏/万曝光」,CSV 口径「涨粉/万曝光」content.window_top_posts/repost_candidates_grade_b:窗口 Top 帖与 B 级(低曝光高收藏,换头重发候选)- pulse 另有
followers_now/follower_trend(粉丝快照趋势);CSV 另有overview.*(周度趋势/净涨粉)
类别关键词表可用 --categories 自定义.json 覆盖(格式:{"类别名": ["关键词", ...]})。
第 3 步:写复盘报告
基于 JSON 写 Markdown 周报,结构固定:
- 本周大盘:曝光、净涨粉、日均净涨粉(对照验收线,如日均 25+)、发帖数;与上周环比。
- 时段结论:哪个北京时间段「涨粉/万曝光」最高。注意剔除单帖爆款扭曲——若某时段数据由 1-2 条爆款贡献,要指出样本量问题,不要直接下结论。
- 类别结论:各类别涨粉效率排名,对照历史基准线,指出漂移。
- 单帖分级:A 级(扩成系列)、B 级(换头重发,列出候选)、C 级(停掉的选题方向)。
- 下周指导:发什么(3-5 个具体选题方向,可对照选题库)、什么时段发、避免什么。
What ships with it
14 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- benchmarks.example.txt 160 B
- README_EN.md 2.7 KB
- README.md 2.4 KB
- scripts/analyze_x_data.py 17 KB runs code
- scripts/fetch_x_api.py 5.3 KB runs code
- scripts/fetch_x_pulse.py 13 KB runs code
- scripts/pulse_to_dashboard.py 7.0 KB runs code
- scripts/tests/test_analyze_x_data.py 2.5 KB runs code
- scripts/tests/test_api_check.py 1.1 KB runs code
- scripts/tests/test_fetch_filter.py 1.1 KB runs code
- scripts/tests/test_pulse_to_dashboard.py 3.2 KB runs code
- scripts/tests/test_topic_feedback.py 1.8 KB runs code
- scripts/topic_feedback.py 4.2 KB runs code
- scripts/track_benchmarks.py 4.5 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 136 lines · 171 tokens per session scan A 20ce9946a6aa
x-content-review is a skill published in the GitHub repository anneheartrecord/charles-skill (3 stars, last pushed 1mo ago), licensed MIT. It adds 171 tokens to every session and 2,439 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
agent-recall
Persistent compounding memory for AI agents. 5 default MCP tools: sessionstart, sessionend, remember, recall, check. Full surface (18 tools) available with --full flag. Two-verb model: inhale (sessionstart) and exhale (sessionend). Correction-first memory with decision trail tracking, watchfor warnings, palace rooms…
mykg
Run mykg knowledge-graph commands inside Claude Code from one slash command /mykg. The user describes intent in natural language (extract, append, sync, resume, approve, walkthrough, parse-docs, fetch-web, query); the skill parses intent, builds the right mykg CLI command from the live --help output, confirms, runs…
swarmvault
Use SwarmVault when the user needs a local-first knowledge vault that writes durable markdown, graph, search, dashboard, review, chat-session, context-pack, task-ledger, static AI export, retrieval, and MCP artifacts to disk from books, notes, transcripts, exports, datasets, slide decks, files, URLs, code, and…
engraph
Index and search document collections using hybrid semantic + graph + full-text search. Use when users need to search knowledge bases, find connections between documents, discover related content via link graphs, or query indexed markdown collections.
pagefly
Connect to your self-hosted PageFly knowledge base. Query documents, search, upload files, read content, and chat with your AI agent — all from Claude Code.
compiler
Scans knowledge/, analyzes documents, generates summaries/concepts/connections to wiki/ with reference graph. Updates existing articles instead of creating duplicates.