← Back to Skills Marketplace
543
Downloads
2
Stars
0
Active Installs
11
Versions
Install in OpenClaw
/install research-to-wechat
Description
A native research-first pipeline that turns a topic, notes, article, URL, or transcript into a sourced article with an evidence ledger, polished Markdown, in...
Usage Guidance
This skill is broadly coherent with its purpose (turn research into a WeChat draft) and includes local Python scripts to fetch, render, upload images, and save drafts. Before installing or providing credentials: 1) Inspect scripts/wechat_delivery.py, scripts/upload-images and scripts/fetch_wechat_article.py to confirm they call only WeChat official endpoints and to see exactly which env vars or flags they use. 2) Do not run the README's curl | bash install line without reading scripts/install-openclaw.sh first. 3) Expect that WECHAT_SECRET (and possibly WECHAT_DRAFT_MEDIA_ID or other tokens) will be needed for L0 automated delivery even though the registry didn't declare them — prefer using assisted- or manual-handoff delivery modes for testing if you don't want to share secrets. 4) If you keep EXTEND.md in your home config, review its content (it can include QR images and URLs). 5) If you need higher assurance, ask the publisher for clarification about the missing declared env vars (WECHAT_SECRET / stable_token) and request a short audit of network endpoints used by the Python scripts.
Capability Analysis
Type: OpenClaw Skill
Name: research-to-wechat
Version: 0.5.5
The 'research-to-wechat' skill bundle is a professional-grade content orchestration pipeline designed to transform various inputs (URLs, PDFs, transcripts) into formatted WeChat articles. The bundle includes a suite of Python scripts (`wechat_delivery.py`, `_wechat_delivery_api.py`, etc.) that handle Markdown-to-HTML rendering with complex design templates and interact with official WeChat APIs using user-provided credentials. An installer script (`install-openclaw.sh`) facilitates setup by downloading assets from a public GitHub repository. The agent instructions in `SKILL.md` and `execution-contract.md` are highly detailed, focusing on content quality, stylistic 'de-AI' checks, and compliance with WeChat's technical constraints. While the skill requires sensitive API keys and performs network operations, all behaviors are transparently documented and strictly aligned with the stated purpose of automated publishing. No evidence of malicious intent, data exfiltration, or unauthorized persistence was found.
Capability Assessment
Purpose & Capability
Name/description align with included scripts (WeChat fetch, render, upload, save-draft). Requiring python3 is appropriate. However the registry metadata lists no required env vars while SKILL.md and the scripts reference WECHAT_APPID and WECHAT_SECRET (and other fields like WECHAT_DRAFT_MEDIA_ID); primaryEnv is set to WECHAT_APPID but WECHAT_SECRET is not declared. This is a packaging/documentation mismatch that should be clarified.
Instruction Scope
SKILL.md limits behavior to research, markdown normalization, WeChat HTML rendering, image upload, and draft save — all consistent with the stated purpose. It also instructs the runtime to read optional author config files (EXTEND.md) from the project dir or the user's home (~/.config/... and ~/.research-to-wechat/) and to read a project AGENTS.md for compliance checks; these are legitimate for customization but mean the skill will read files from the user's home and project. The instructions also mention using browser tools and a Pencil MCP server when available. There is no explicit instruction in SKILL.md to read unrelated secrets like SSH keys, but you should expect the skill to read the declared config paths.
Install Mechanism
There is no automated install specification in the registry (instruction-only), which is lower risk. The repository includes an install-openclaw.sh and the README suggests a curl | bash install command; that is a convenient but higher-risk pattern — running the install script without reviewing it is not recommended. The included Python scripts are local and not pulled from arbitrary external URLs at runtime according to the provided docs.
Credentials
The skill’s primary credential is WECHAT_APPID (expected). But runtime steps (L0 delivery, upload-images, save-draft) require WECHAT_SECRET (and possibly WECHAT_DRAFT_MEDIA_ID) to obtain access_token and call WeChat APIs; WECHAT_SECRET is referenced in README/SKILL.md but is not declared as a required env var in the registry. The capability map also mentions an `access_token` from a `stable_token` variable which is unexplained. This mismatch between declared and actually-used credentials is a proportionality/packaging issue and could lead users to accidentally supply keys where the skill expects different values or to be asked for secrets at runtime unexpectedly.
Persistence & Privilege
The skill is not marked always:true and does not request system-wide persistence. It writes workspace artifacts to a per-article directory and reads optional user-level EXTEND.md/project AGENTS.md as documented. There is no evidence it attempts to modify other skills or system-wide agent settings.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install research-to-wechat - After installation, invoke the skill by name or use
/research-to-wechat - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.5.5
Add pre-delivery compliance gate (force re-read AGENTS before Phase 8), document 7 known renderer issues with detection and fix methods
v0.5.4
Checklist #2 now greps for all negation forms (不是/不只/不再/而非/而是) individually instead of only paired patterns. Phase 4 adds source attribution rules: forum content is topic inspiration only, no usernames, no event narration, no numbered lists.
v0.5.3
Add mandatory project AGENTS.md compliance verification step before delivery, with checklist item #21 that walks through every rule section
v0.5.2
Add 6 new pre-draft checks: upload-map must include cover+QR, reference section text-align:left, no numbered lists, no forum usernames, bare wechatqr check, HTML rendering rules for refs and lists
v0.5.1
Add pre-draft checklist (15 checks) and mandatory --cover-image param
v0.6.0
Version 0.6.0
- Introduced a native WeChat delivery pipeline: new core scripts handle rendering, image upload, and draft creation (no longer dependent on external/legacy skills).
- Added core delivery scripts: `wechat_delivery.py` and supporting modules for `render`, `upload-images`, `save-draft`, and `design-catalog` operations.
- New author configuration system: workflow can now append author-specific CTA blocks using `EXTEND.md` (see `author-config.md` for details).
- Updated compliance and normalization steps: integrated machine-verifiable Chinese de-AI checklist and WeChat content compliance scanning.
- Expanded contract and documentation: design catalog, capability map, and delivery logic now split into dedicated, referenced files for clarity.
- Improved asset and CDN image upload handling: skips re-uploading unchanged assets, merges new URLs, and avoids redundant CTA QR code uploads.
v0.4.2
v0.4.2: PDF figure extraction (auto-extract charts/tables/diagrams from papers as source images), source figure priority over AI-generated images, WeChat HTML compatibility rules (inline CSS, section not div, no flex/grid, dark background wrapping), new wechat-compat.md reference with API endpoints and CDP workflow
v0.4.1
v0.4.1: Add WeChat HTML compatibility constraints (inline CSS, use section, background
requirements) and detailed WeChat API draft/upload operational manual (CDP base64 image injects, AppMsgId
capitalization fixes) from debugging sessions.
v0.4.0
Writing frameworks (deep-analysis + tutorial), Pencil design templates, md2wechat HTML converter, cover spec, multi-platform distribution, OpenClaw compatibility
v0.1.1
Fix portability: replace absolute local paths in SKILL.md references with relative links for cross-machine installs.
v0.1.0
Initial release: topic-to-draft WeChat article orchestration with style engine, capability aliases, and phased execution contract.
Metadata
Frequently Asked Questions
What is Research To Wechat?
A native research-first pipeline that turns a topic, notes, article, URL, or transcript into a sourced article with an evidence ledger, polished Markdown, in... It is an AI Agent Skill for Claude Code / OpenClaw, with 543 downloads so far.
How do I install Research To Wechat?
Run "/install research-to-wechat" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Research To Wechat free?
Yes, Research To Wechat is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Research To Wechat support?
Research To Wechat is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Research To Wechat?
It is built and maintained by clarezoe (@clarezoe); the current version is v0.5.5.
More Skills