计划执行与检查节点
executing-plans
按既定实施计划完成工作,保持过程检查和反馈。
oschina【效率与协作】Read the connected 开源中国 / OSChina (my.oschina.net) account and create Markdown blog drafts with the user's own login cookies (BYOC). Use when the user mentions 开源中国 / OSChina, wants to save a 开源中国 draft, or asks who their connected 开源中国 account is.
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Drives the user's real 开源中国 account through the same
apiv1.oschina.net endpoints the site uses, authenticated by the login cookie
they captured with the ACE extension. No browser, no third-party deps — just
urllib.
The connector injects the cookie jar as a JSON env var $OSCHINA_COOKIES.
Never print it.
python3 "$SKILL_DIR/scripts/oschina.py" whoami
If $SKILL_DIR points at a different skill loaded in the same turn, resolve
this skill's directory explicitly before running the commands below.
开源中国 has no public publish API. Every open-source client for this platform (including the upstream adapter this skill is modelled on) stops at draft creation. This skill does the same: it saves the Markdown as a draft and returns the editor URL. Tell the user plainly that they must open that URL and click publish themselves — do not claim the article was published.
python3 "$SKILL_DIR/scripts/oschina.py" whoami
If this fails with an auth error, the cookie has expired. Ask the user to
reconnect at https://studio.acedata.cloud/console/connectors rather than retrying.
Prepare the complete Markdown in a file. The first call is always a dry run and does not write anything.
# Dry run — shows exactly what would be written.
python3 "$SKILL_DIR/scripts/oschina.py" draft \
--title "标题" --content-file /tmp/article.md
# Actually create the draft after the user confirms.
python3 "$SKILL_DIR/scripts/oschina.py" draft \
--title "标题" --content-file /tmp/article.md --confirm
--confirm is valid only as the final argument. Show the title and full
content to the user before writing.
contentType: 1). Do not pre-render it to HTML.--catalog takes a 开源中国 catalog id; the default 0 uses the account's
default catalog. If the API rejects the catalog, ask the user for the right
id rather than guessing.This skill only produces drafts, so do not call publish_artifact. Report
the returned draft_id and edit_url to the user instead.