Skip to content

staruhub/ClaudeSkills / skills/Geek-skills-wechat-article-writer

wechat-article-writer

Verified

专业微信公众号文章助手,支持四个独立且可组合模式:article 写正文;image-prompts 从文章生成版本化、provider-neutral 的图片提示词 manifest 与稳定占位符,但不调用生图;layout 把文章和 manifest 确定性转换为微信安全的内联 HTML;full-pipeline 串联正文、提示词/占位符和排版。用户提到公众号、微信文章、推文、配图提示词、图片 manifest、公众号排版、内联 HTML 或完整图文流水线时触发。支持企业官号、个人技术博客、活动回顾和产品评测。不用于:带引用的研究报告(用 deep-research)、PPT/组图(用 deck-studio)、英文文章、仅润色一小段。永不自动发布,永不把提示词或占位符声称为已生成图片。

references/PIPELINE-CONTRACT.md

View on GitHub
# WeChat article pipeline contract

This contract keeps writing, image prompting, image generation, layout, and
publishing as separate capabilities. The skill implements the first, second,
and layout handoff only. It never publishes.

## Modes

### `article`

Input: source material and writing brief.

Output: UTF-8 `article.md`. Do not create a manifest unless the user requests
image prompts or the full pipeline.

### `image-prompts`

Input: final or draft `article.md`.

Output:

1. the same article with deterministic image anchor/placeholder pairs;
2. `image-manifest.json` conforming to
   `schemas/image-manifest.schema.json`.

Stop after prompt generation. Set `provider_status` to `prompt-only`. Do not
call an image provider and do not claim an image exists.

### `layout`

Input: `article.md` and, when placeholders exist,
`image-manifest.json`.

Output: a single `<section>` fragment in `layout.html` using inline styles.
Every manifest item becomes one traceable placeholder block unless a verified
generated asset was explicitly supplied.

### `full-pipeline`

Run `article` → `image-prompts` → `layout`. Emit the three artifacts and stop.
This mode still does not generate images or publish the article.

## Deterministic image mapping

For an image ID `img-agent-contract`:

- anchor: `<!-- image:img-agent-contract -->`
- placeholder: `{{IMAGE:img-agent-contract}}`

Place the two lines together at the intended insertion point. The manifest must
repeat both strings exactly. IDs are stable across reruns when purpose and
placement have not changed. Use semantic IDs; never use array indexes or random
UUIDs.

`article_sha256` is calculated over canonical article text bytes: decode as
strict UTF-8, replace CRLF and bare CR with LF, encode as UTF-8, then hash with
SHA-256. No other Unicode or whitespace normalization is allowed. Mapping is
then checked on that canonical text: the anchor line must be immediately
followed by its placeholder line, with exactly one LF separator and no blank
line. This gives Windows and Unix the same digest without weakening adjacency.

The manifest validator rejects:

- duplicate or unstable IDs;
- missing fields or dimensions;
- aspect ratio inconsistent with width/height;
- anchor/placeholder strings that do not derive from the ID;
- article/manifest hash mismatch;
- missing, duplicate, or extra article mappings;
- `generated` status without an explicit asset reference.

## Provider-neutral boundary

`positive_prompt` and `negative_prompt` describe the desired result without
provider-specific flags, model names, account IDs, API parameters, or secrets.
`provider_status` records lifecycle only:

- `not-requested`
- `prompt-only`
- `generating`
- `generated`
- `failed`
- `blocked`

The skill itself may emit only `not-requested` or `prompt-only`. A separate image
provider adapter may update status and `asset_ref` after real execution and
verification.

## Layout Skill handoff

Preferred host interface:

- capability: `排版输出` / WeChat inline layout;
- inputs: exact `article.md`, optional validated manifest, theme ID or default;
- outputs: `layout.html`, consumed image IDs, unresolved image IDs, warnings;
- required behavior: preserve mapping, escape raw HTML, use inline styles only,
  enforce mobile width, and perform no publishing.

Do not couple this prompt skill to a desktop UI, clipboard, image host, account,
or Tauri command. If the host layout capability is absent, run:

`python3 scripts/render_wechat_layout.py --article article.md --manifest image-manifest.json --output layout.html`

The fallback renderer is deliberately small and deterministic. It supports the
core Markdown needed by the example and safely escapes unsupported raw HTML.

## Failure behavior

- Missing manifest with article placeholders: stop and request or rebuild the
  manifest.
- Invalid manifest: show the validator errors; do not render partial HTML.
- Layout adapter unavailable: use the deterministic fallback and disclose the
  reduced Markdown feature set.
- Image provider unavailable or denied: keep prompt-only placeholders.
- Publication request: return files and say publication requires a separate,
  explicit user-controlled action.

## Validation

Run:

`python3 scripts/validate_image_manifest.py image-manifest.json --article article.md`

Then render. Treat any non-zero exit as a hard failure.
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 322kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall