ClawHub

Wechat Publisher DraftGet

Orchestrate the official WeChat draft publishing path for already-prepared article HTML or Markdown using md2wechat, with mandatory draft/get verification before reporting success.

Audits

Pass
ClawScanReview

Agentic behavior and permission review.

Static analysisPass

Pattern checks against bundled files.

VirusTotalPass

Multi-engine malware detections and file reputation.

Install

openclaw skills install wechat-publisher-md2wechat-draftget

wechat-publisher-md2wechat-draftget

Use wechat-publisher-md2wechat-draftget when the user wants to:

  • push a prepared WeChat article into the Official Account draft box
  • turn Markdown into WeChat-compatible HTML and then create a draft
  • use a stable publishing wrapper instead of manually chaining upload_image -> create_draft -> draft/get
  • verify that a draft really exists in the WeChat backend before reporting success
  • troubleshoot the current md2wechat create_draft publishing chain

What This Skill Is For

This skill is the publishing execution layer for WeChat Official Account drafts.

It is responsible for:

  • config validation
  • Markdown → HTML conversion when needed
  • cover upload
  • draft.json generation
  • md2wechat create_draft
  • WeChat draft/get backend verification
  • outputting machine-readable verification artifacts

It is not responsible for:

  • topic selection
  • long-form writing
  • AI-trace review
  • editorial approval
  • layout strategy decisions before HTML exists
  • archive / final content governance

Those stay in the main production workflow.

Canonical Publish Path

Prepared article HTML or Markdown
→ md2wechat config validate
→ inspect / preview / convert (if Markdown)
→ md2wechat upload_image cover.*
→ generate draft.json
→ md2wechat create_draft draft.json
→ call WeChat draft/get
→ verify title / author / cover / content / visibility

Hard Rules

  1. md2wechat create_draft is the formal draft creation action.
  2. test-draft is only for chain testing, not formal publishing.
  3. convert --mode ai returning CONVERT_AI_REQUEST_READY does not mean a draft was created.
  4. AI-mode output must be materialized into real WeChat-compatible HTML before entering this skill.
  5. Command success is not task success.
  6. Final success requires backend verification.
  7. Preferred execution path is scripts/publish.sh, not an ad-hoc hand-built sequence.
  8. Local preview and local convert artifacts are not WeChat backend proof.
  9. If create_draft succeeds but draft/get fails, the result must be treated as unverified or blocked.
  10. media_id alone is only a submission receipt, not a success proof.

Inputs

Preferred prepared package:

wechat_package/
  article.html
  cover.png
  images/
  publish-checklist.md

Also supported:

  • article.md + cover image
  • article.html + cover image

Requirements

The following must already be available:

  • md2wechat
  • python3

Config file:

~/.config/md2wechat/config.yaml

Expected WeChat credentials:

wechat:
  appid: "公众号 AppID"
  secret: "公众号 AppSecret"

The current public IP must also be in the WeChat Official Account IP allowlist.

Validate Before Publish

md2wechat config validate --json

Standard Execution

Primary script:

skills/wechat-publisher/scripts/publish.sh

Usage:

./scripts/publish.sh <article.md|article.html> <cover-image> [title] [author] [digest]

Examples:

./scripts/publish.sh article.html cover.png "现在找工作，最累的不是被拒" "野哥" "找工作最累的不是被拒，而是一直没有回音。"
./scripts/publish.sh article.md cover.png "标题" "野哥" "摘要"

What The Script Does

  1. md2wechat config validate --json
  2. If input is Markdown:
    • md2wechat inspect ... --json
    • md2wechat preview ... --json
    • md2wechat convert ... --output ... --json
  3. md2wechat upload_image cover.png --json
  4. Generate *.md2wechat-create-draft.json
  5. md2wechat create_draft *.json --json
  6. Call official WeChat draft/get
  7. Save verification output

Output Artifacts

The script writes artifacts next to the article:

*.md2wechat-inspect.json
*.md2wechat-preview.json
*.md2wechat-convert.json
*.md2wechat-cover-upload.json
*.md2wechat-create-draft.json
*.md2wechat-create-draft-result.json
*.md2wechat-create-draft-verify.json

The final source of truth is:

*.md2wechat-create-draft-verify.json

Success Criteria

Only report draft push success when both are true:

  1. md2wechat create_draft returned success
  2. draft/get verification passed and the backend draft is really present

Otherwise classify as one of:

  • submitted but unverified
  • verification failed
  • blocked by config / credential / IP / content issue

Draft JSON Shape

{
  "articles": [
    {
      "title": "标题",
      "author": "野哥",
      "digest": "摘要",
      "content": "<section style=\"...\">正文 HTML</section>",
      "thumb_media_id": "封面素材 media_id",
      "show_cover_pic": 0
    }
  ]
}

Constraints:

  • articles must not be empty
  • title required, recommended ≤ 32 chars
  • author recommended ≤ 16 chars
  • digest recommended ≤ 120–128 chars
  • content required and must be WeChat-compatible HTML
  • thumb_media_id must come from md2wechat upload_image or equivalent WeChat material upload

Verification Checklist

At minimum verify:

  • draft really exists
  • title matches expected value
  • author matches expected value
  • thumb_media_id exists
  • content exists
  • no leaked local paths such as /home/ye/ or content-factory/
  • content has inline styles or otherwise clearly matches publish-ready WeChat HTML
  • article images are actually present when the article uses body images
  • for Markdown input, the process did not stop at local preview/convert only
  • for AI mode upstream flows, the published content is real generated HTML, not the request-ready payload

Common Failure Modes

Invalid IP

errcode=40164 invalid ip ... not in whitelist

Action: add the current public IP to the WeChat Official Account backend allowlist.

Invalid AppSecret

errcode=40125 invalid appsecret

Action: fix ~/.config/md2wechat/config.yaml.

AI mode only returned a request

{
  "code": "CONVERT_AI_REQUEST_READY",
  "status": "action_required"
}

This is not success. It only means the AI conversion request is ready. Continue by producing HTML, then run the draft chain.

Operational Guidance For Agents

  • Use this skill only when draft creation is actually requested.
  • Prefer this skill over improvising raw shell steps.
  • Never claim success from media_id, preview, or convert alone.
  • Treat verification as mandatory, not optional polish.
  • If verification fails, report the exact blocker instead of vague success language.

Related Files

  • README.md — quick operator-facing usage
  • references/troubleshooting.md — failure handling
  • references/themes.md — related publishing notes
  • scripts/publish.sh — canonical execution wrapper
  • scripts/setup.sh — environment setup helper

Experience Captured

2026-05-14

Verified working path using an older article:

AI-ready HTML
→ md2wechat upload_image
→ draft.json
→ md2wechat create_draft
→ WeChat draft/get verification passed

Verified fields included title, author, digest, cover, content, inline styles, and no local path leakage.

2026-05-15

Key lessons added to the standard skill:

  • formal publishing should default to scripts/publish.sh
  • media_id is only a receipt
  • CONVERT_AI_REQUEST_READY is an action-required state, not a success state
  • preview / convert / test-draft can mislead people into thinking the draft is already in the backend
  • if backend draft verification fails, treat the run as unverified or blocked instead of assuming eventual consistency