source-onboarding
Create or repair Askme website discovery skills (*-discovery / *-platform-discovery).
Skill by paxionfull
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/paxionfull-source-onboarding
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing, productivity
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/source-onboarding
What's inside
- Skill name
source-onboarding- Description
- Create or repair Askme website discovery skills (*-discovery / *-platform-discovery). ALWAYS use this skill when onboarding a new site from a URL, implementing WebsiteFeedAdapter, writing discover.py / source.yaml / SKILL.md, fixing discovery_validate or refresh failures, login/cookie ASKME_AUTH_REQUIRED, or http_client / hints / list_index rules. Prefer this over any existing *-discovery skill for the onboarding workflow itself; pick structural references from the dynamic Discovery skill catalog (name/description/path). Not for day-to-day fetching of an already-onboarded source.
- License
- MIT
- Instructions
- 86 lines in SKILL.md · view on GitHub
Before installing
- Instructions call the network
Agent Skills spec issues
- files: referenced but missing: scripts/discover.py
Files (2)
- CONTRACT.md
- SKILL.md
SKILL.md (first lines)
# 数据源接入(source-onboarding)
Askme 为**具体网站**编写 discovery skill(Askme Agent / Cursor Agent),不是 RSS,不是模板填空。
**动手前必读契约:** [CONTRACT.md](CONTRACT.md)(接口、HTTP、hints、登录墙)。违反契约则 pipeline 无法复用。
## 原则
1. 用户通常只提供网站链接;名称与 slug 由系统从域名推导(任务 prompt 会给出最终 slug)
2. 先看网站证据(入口页、API 样本),再写专属代码;URL/字段/Referer/分页必须来自该站
3. **参考已有 discovery skill(强制)**:只从动态目录 [DISCOVERY_SKILL_CATALOG.md](../../discovery/_lib/DISCOVERY_SKILL_CATALOG.md)(或任务 prompt 内嵌的同表)按 **name / description** 选 ≥2 个样例;学结构与 `_lib` 用法,**不可照搬** URL/字段
4. **禁止 RSS/Atom**
5. 验证失败则自行修复并重跑 `discovery_validate.py`
## 如何选参考 skill(强制,至少 2 个)
仓库内已有站级 / 平台级 discovery(含内置与用户接入的)都是合法参考;**选型只认 catalog 表**,不要另找 playbook,也**禁止** `ls`/glob 全目录碰运气。
1. 打开任务 prompt 中的「Discovery skill 参考目录」,或读取 `skills/discovery/_lib/DISCOVERY_SKILL_CATALOG.md`
2. 根据侦察结果(JSON API / HTML 列表 / sitemap / 需 Cookie 等),用表中的 **name + description** 选出 **至少 2 个**形态最接近的 skill
3. 若目标站已有同站或同平台 skill,优先列入参考
4. 只打开所选 skill 的 `path` 下 `scripts/discover.py`、`source.yaml`(必要时 `SKILL.md`),只学:
- WebsiteFeedAdapter 组织方式
- `http_client` / `detail_hints` / `list_index` / `auth_cookie` 用法
5. **禁止**复制其 API URL、字段路径、Referer、假数据;字段必须来自目标站证据
6. 动手写代码前,先明确写出所选参考 skill 的 name 与 path
## 未知站 — 任务步骤
1. 用 curl / Python 侦察真实列表与正文 API
2. 阅读 [CONTRACT.md](CONTRACT.md)
3. 按上节从 catalog 选 ≥2 个参考 skill 并阅读
4. 创建 `skills/discovery/{slug}-discovery/`:
- `scripts/discover.py`(完整 WebsiteFeedAdapter)
- `source.yaml`、`SKILL.md`(description 只描述本站与抓取方式;勿写成通用接入指南,勿加 “Not for onboarding…”)
5. 运行:`python skills/discovery/_lib/discovery_validate.py {slug}`
6. 失败则按报错修到通过;完成后一句话说明 `feed_id`、所选参考 skill、与验证结果(若停在 `ASKME_AUTH_REQUIRED`,写出 slot)
IDE 手动接入同样遵循上述步骤;完成后可 `POST /api/feeds/reload-skills`。
## 修复已有 discovery skill
1. 先加载本 skill + [CONTRACT.md](CONTRACT.md),**不得破坏契约**
2. 按用户反馈 / 验证报错定位 `discover.py` / `source.yaml`(平台 skill 优先 `_lib/{platform}_common.py`)
3. 必要时从 catalog 再打开 1~2 个相近 `*-discovery` 对照实现,仍禁止照搬 URL/字段
4. 站级:只改 `skills/discovery/{slug}-discovery/`;禁止改 slug、`FEED_ID`、删目录
5. 平台级(slug 以 `-platform` 结尾):可改平台 skill 与对应 `_lib`;禁止新建 per-account skill
6. 跑通 `discovery_validate.py`
## 已知多账号平台(系统侧确定性接入)
部分多账号平台由 backend `detect_platform` + `platform_registry` 走脚手架,**不**为每个账号新建目录:
```text
detect → 缺 Cookie 则 ASKME_AUTH_REQUIRED → probe → platform_accounts + 平台 skill → validate
```
- 一平台一个 skill:`skills/discovery/{platform}-platform-discovery/`
- 同构:`SKILL.md` + `source.yaml`(`platform_skill: true`)+ 真实 `discover.py`(`require_account()` + `BoundPlatformAdapter`)
- 缺 Cookie **不**触发 auto_repair;须用户授权
平台 skill 文件要求与运行时约定亦须满足 [CONTRACT.md](CONTRACT.md) 的接口/HTTP/鉴权规则。具体平台清单以 `platform_registry` / catalog 为准,勿在本文硬编码罗列。
微博客 / 论坛类平台:**只抓作者原创主帖**,不抓回复、评论、RT、QT(详见 CONTRACT「列表范围」);实现落在平台 `discover.py`,勿为每个账号单独开例外。
## 自动化与 API(产品说明)
需配置**对话模型 API Key**(设置页 LLM,Askme Agent)或 **Cursor API Key**(Cursor SDK)。优先使用已配置的用户 LLM。
接入任务 prompt 会注入动态 catalog(并刷新 `_lib/DISCOVERY_SKILL_CATALOG.md`)。
```text
resolve_engine(askme|cursor) → agent(Askme:recon→select→implement→validate)
→ discovery_validate → reload feeds
```
Askme Agent 与 Cursor 同一收口:写完整 skill → `discovery_validate`;
撞 `ASKME_AUTH_REQUIRED` 则为半成功(`needs_auth`,等 Cookie),禁止只写 yaml 早停。
失败且 `auto_repair` 时按报错迭代修平台/站级 skill;`ASKME_AUTH_REQUIRED` 不走自动修复。
- `POST /api/sources/onboard`(SSE)
- 日志:`data/onboarding-logs/{job_id}.jsonl`
- 后端入口:`backend/onboarding/source_onboarding_cursor.py`、`onboarding/agent/`、`source_platform_onboard.py`、`platform_registry.py`From paxionfull/askme, MIT.
More from paxionfull/askme
- 0.529score
zhihu-platform-discovery
paxionfull
Discovers Zhihu user/org articles via official API.
SkillBFreshness: BClaude Code +9 - 0.529score
jiqizhixin-discovery
paxionfull
Discovers articles from 机器之心 (jiqizhixin.com) via the official article library JSON API.
SkillBFreshness: BClaude Code +9 - 0.529score
x-platform-discovery
paxionfull
Discovers X (Twitter) user original posts only (no replies, retweets, or quote tweets) via logged-in GraphQL (Cookie auth_token+ct0), with FxEmbed /…
SkillBFreshness: BClaude Code +9 - 0.528score
qbitai-discovery
paxionfull
Discovers articles from 量子位 (qbitai.com) via WordPress REST API.
SkillBFreshness: BClaude Code +9 - 0.526score
reddit-platform-discovery
paxionfull
Discovers Reddit subreddit posts via arctic-shift / pullpush public archive (optional login Cookie for www.reddit.com JSON).
SkillBFreshness: BClaude Code +9 - 0.510score
chat-rag
paxionfull
对话时的角色 system prompt(问答与选定文章摘要共用;引用硬性规则由系统在问答时自动追加)
SkillBFreshness: BClaude Code +9 - 0.504score
impeccable
paxionfull
Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or…
SkillBFreshness: BClaude Code +9 - 0.280score
AI_news
paxionfull
结构化整理规则
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Aug 17, 2026 (2 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/paxionfull/askme
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from paxionfull
- 0.529score
x-platform-discovery
paxionfull
Discovers X (Twitter) user original posts only (no replies, retweets, or quote tweets) via logged-in GraphQL (Cookie auth_token+ct0), with FxEmbed /…
SkillBFreshness: BClaude Code +9 - 0.529score
zhihu-platform-discovery
paxionfull
Discovers Zhihu user/org articles via official API.
SkillBFreshness: BClaude Code +9 - 0.529score
jiqizhixin-discovery
paxionfull
Discovers articles from 机器之心 (jiqizhixin.com) via the official article library JSON API.
SkillBFreshness: BClaude Code +9 - 0.528score
qbitai-discovery
paxionfull
Discovers articles from 量子位 (qbitai.com) via WordPress REST API.
SkillBFreshness: BClaude Code +9 - 0.526score
reddit-platform-discovery
paxionfull
Discovers Reddit subreddit posts via arctic-shift / pullpush public archive (optional login Cookie for www.reddit.com JSON).
SkillBFreshness: BClaude Code +9 - 0.510score
chat-rag
paxionfull
对话时的角色 system prompt(问答与选定文章摘要共用;引用硬性规则由系统在问答时自动追加)
SkillBFreshness: BClaude Code +9
Cite this page
source-onboarding skill. Agent Mart. https://agentmart.dev/skills/paxionfull-source-onboarding (last updated 2026-10-24)