← 返回 Skills 市场
bibaofeng

last30days

作者 bibaofeng · GitHub ↗ · v1.0.4 · MIT-0
cross-platform ✓ 安全检测通过
119
总下载
0
收藏
0
当前安装
3
版本数
在 OpenClaw 中安装
/install last30days-aisa
功能描述
Research the last 30 days across Reddit, X/Twitter, YouTube, TikTok, Instagram, Hacker News, Polymarket, GitHub, and web search. Use when: you need recent so...
使用说明 (SKILL.md)

\r \r

last30days\r

\r Research recent evidence across social platforms, community forums, prediction markets, GitHub, and grounded web results, then merge everything into one brief.\r \r

When to use\r

\r

  • Use when you need a last-30-days research brief on a person, company, product, market, tool, or trend.\r
  • Use when you want a recent competitor comparison, launch reaction summary, creator/community sentiment scan, or shipping update.\r
  • Use when you want structured JSON with query_plan, ranked_candidates, clusters, and items_by_source.\r \r

When NOT to use\r

\r

  • Do not use for timeless encyclopedia questions with no recent evidence requirement.\r
  • Do not use when you need only one official source and do not want social/community signals.\r \r

Capabilities\r

\r

  • AISA-hosted planning, reranking, synthesis, grounded web search, X/Twitter search, YouTube search, and Polymarket search.\r
  • Public Reddit and Hacker News retrieval with fail-soft behavior.\r
  • Official GitHub API search when GH_TOKEN or GITHUB_TOKEN is available.\r
  • Hosted discovery for TikTok, Instagram, Threads, and Pinterest when enabled in runtime config.\r \r

Setup\r

\r

  • AISA_API_KEY is the main hosted credential.\r
  • GH_TOKEN or GITHUB_TOKEN is optional for GitHub search only.\r
  • Python 3.12+ is required.\r \r
for py in /usr/local/python3.12/bin/python3.12 python3.14 python3.13 python3.12 python3; do\r
  command -v "$py" >/dev/null 2>&1 || continue\r
  "$py" -c 'import sys; raise SystemExit(0 if sys.version_info >= (3, 12) else 1)' || continue\r
  LAST30DAYS_PYTHON="$py"\r
  break\r
done\r
```\r
\r
## Quick Reference\r
\r
```bash\r
bash "${SKILL_ROOT}/scripts/run-last30days.sh" "$ARGUMENTS" --emit=compact\r
"${LAST30DAYS_PYTHON}" "${SKILL_ROOT}/scripts/last30days.py" "$ARGUMENTS" --emit=json\r
"${LAST30DAYS_PYTHON}" "${SKILL_ROOT}/scripts/last30days.py" "$ARGUMENTS" --quick\r
"${LAST30DAYS_PYTHON}" "${SKILL_ROOT}/scripts/last30days.py" "$ARGUMENTS" --deep\r
"${LAST30DAYS_PYTHON}" "${SKILL_ROOT}/scripts/last30days.py" "$ARGUMENTS" --search=reddit,x,grounding\r
"${LAST30DAYS_PYTHON}" "${SKILL_ROOT}/scripts/last30days.py" --diagnose\r
```\r
\r
## Inputs And Outputs\r
\r
- Input: a topic or comparison query such as `OpenAI Agents SDK`, `OpenClaw vs Codex`, or `Peter Steinberger`.\r
- Output: synthesized research plus `provider_runtime`, `query_plan`, `ranked_candidates`, `clusters`, and `items_by_source`.\r
\r
## Example Queries\r
\r
- `last30days OpenAI Agents SDK`\r
- `last30days Peter Steinberger`\r
- `last30days OpenClaw vs Codex`\r
- `last30days Kanye West --quick`\r
\r
安全使用建议
This skill appears to do what it says: it sends queries and content to the AISA-hosted services (so prompts and retrieved content travel to that provider) and it will save findings locally (a SQLite DB and brief JSON files). Before installing: (1) decide whether you are comfortable sending queries/data to the AISA service and consider using a scoped/limited API key; (2) if you prefer to control local storage, set LAST30DAYS_DATA_DIR and LAST30DAYS_CONFIG_DIR to a directory you manage or run the skill in an isolated container; (3) only provide GH_TOKEN/GITHUB_TOKEN if you want GitHub results; (4) review the scripts (they are included in the bundle) if you need to confirm what is stored or transmitted; (5) monitor network activity the first time you run it if you want extra assurance. Overall the skill is internally consistent with its stated purpose.
功能分析
Type: OpenClaw Skill Name: last30days-aisa Version: 1.0.4 The last30days skill is a comprehensive research tool designed to aggregate and synthesize data from social platforms (X, Reddit, TikTok, Instagram), GitHub, and web search. The code is well-structured, using standard Python libraries and legitimate APIs (GitHub, Reddit, and the AIsa backend at api.aisa.one). It demonstrates good security practices, such as checking file permissions for secrets in scripts/lib/env.py and using XML-style fencing (<untrusted_content>) in scripts/lib/rerank.py to mitigate prompt injection risks when processing search results. The persistence mechanism (SQLite) and webhook delivery features in scripts/store.py and scripts/watchlist.py are functionally aligned with the tool's purpose as a monitoring and briefing agent.
能力标签
cryptocan-make-purchases
能力评估
Purpose & Capability
Name/description (recent multi-source research) matches the code and declared requirements: AISA_API_KEY for hosted planning/search, python3 and bash to run the included scripts. The repository includes providers for Reddit, X, YouTube, TikTok, Instagram, Hacker News, Polymarket, GitHub, and web grounding as advertised.
Instruction Scope
Runtime instructions and scripts perform network calls to AISA and various provider APIs, synthesize/rerank results, and persist outputs. The skill writes data (SQLite DB and JSON briefs) to disk (defaults: a .last30days-data directory relative to the run directory and ~/.local/share/last30days/briefs for briefings). SKILL.md does not instruct reading unrelated secrets, but code will read optional env vars like GH_TOKEN/GITHUB_TOKEN when present. This read/emit behavior is expected for a research aggregator but is worth knowing.
Install Mechanism
No install spec; this is an instruction/script bundle with no remote downloads. All code ships in the bundle, so nothing is fetched from arbitrary URLs during install. Risk from install-phase is low.
Credentials
The skill requires a single primary credential (AISA_API_KEY) which is justified by use of AISA-hosted planning/search. GH_TOKEN/GITHUB_TOKEN are optional and used only for GitHub API access. No unrelated cloud or secret credentials (AWS, etc.) are required by default.
Persistence & Privilege
The skill persists research to local storage (SQLite DB and saved briefs) and creates directories under the run directory and in the user's home (~/.local/share). It does not request 'always: true' or modify other skills' configs. If you run it, expect local files and a DB; you can control locations with LAST30DAYS_DATA_DIR / LAST30DAYS_CONFIG_DIR environment variables.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install last30days-aisa
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /last30days-aisa 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.4
- Updated version to 1.0.4. - Changed homepage and repository URLs to https://github.com/AIsa-team/agent-skills. - Removed the sample file .codex-plugin/plugin.json. - Minor edit to the argument hint in metadata.
v1.0.2
Version 1.0.2 – Major update with refactor and documentation improvements - Reorganized and expanded codebase: 37 files added (including .pyc modules), 9 legacy/source-specific files removed. - SKILL.md overhauled: clearer audience guidance (when/when not to use), improved setup instructions, concise commands, and capability descriptions. - Enhanced support for AISA-hosted planning, reranking, clustering, synthesis, and multi-platform search pipelines. - Now explicitly lists required/optional credentials and tools (AISA_API_KEY, GH_TOKEN, Python 3.12+). - Output schema and invocation patterns clarified for end users.
v1.0.0
last30days 1.0.0 — Initial release, AISA-first multi-source research - Supports research and social synthesis from Reddit, X/Twitter, YouTube, TikTok, Instagram, Hacker News, Polymarket, GitHub, and the web, fully routed via AIsa API. - New Python 3.12+ runtime and CLI with flexible search modes (quick, standard, deep, or custom sources). - Single API key (`AISA_API_KEY`) setup—no browser cookie scanning or legacy credential flows required. - Adds watchlist tracking, briefing generation, and search quality evaluation scripts. - Concise, evidence-weighted reporting and synthesis across cross-platform sources.
元数据
Slug last30days-aisa
版本 1.0.4
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 3
常见问题

last30days 是什么?

Research the last 30 days across Reddit, X/Twitter, YouTube, TikTok, Instagram, Hacker News, Polymarket, GitHub, and web search. Use when: you need recent so... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 119 次。

如何安装 last30days?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install last30days-aisa」即可一键安装,无需额外配置。

last30days 是免费的吗?

是的,last30days 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

last30days 支持哪些平台?

last30days 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 last30days?

由 bibaofeng(@bibaofeng)开发并维护,当前版本 v1.0.4。

💬 留言讨论