Generates WeChat article briefs, evidence-backed claims, and styled first drafts from a confirmed topic.
npx skills add https://github.com/oaker-io/wewrite --skill wewrite-write
Please help me install this Agent Skill. Skill: wewrite-write Source: https://github.com/oaker-io/wewrite/tree/main/skills/wewrite-write Command: npx skills add https://github.com/oaker-io/wewrite --skill wewrite-write
Writing a WeChat official account article that is both factually grounded and stylistically consistent requires juggling topic briefs, source verification, persona voice, and draft structure. This Skill turns a confirmed topic into a complete article task brief, a claims-and-evidence ledger, and a first draft, without fabricating personal experiences or unsupported facts.
brief.yaml, then records every fact, inference, and opinion in claims.yaml with verified source references.Ask the agent to write the article body for the currently selected WeChat topic using the wewrite-write module.
High-intent search queries and answers about installing and using this skill.
Run the write module after a topic is confirmed, or provide a topic directly and it creates a draft-mode task. It produces an article brief, a claims ledger with verified sources, and a first draft saved as task artifacts.
Select one of the seven built-in personas such as midnight-friend, cold-analyst, or tech-coder, or add a custom persona YAML file. Personas control tone, rhythm, and uncertainty expression but never override facts or evidence.
No. Personal experiences are only used when the user explicitly supplies them in the current task. When no personal material is available, the draft uses observations, questions, or judgments as openings instead of invented anecdotes.
The module degrades gracefully by writing only analysis that does not depend on current data, removing unverifiable numbers and quotes, and recording the degradation in the task. Unsupported claims never enter the draft.
Yes. Setting the use_writer_model flag sends the brief, claims, source ledger, and style constraints to a dedicated writing model via the wewrite llm-write command. If that call fails, the current agent writes the draft directly.
SpaceZephyrGenerate multi-angle topic ideas from articles, tweets, and social media content.
danielbaustinGenerate title options, outline, draft, and editorial notes from an article brief.
WingedGuardianGenerate LinkedIn posts in the user's authentic voice from voice exemplars.
jackatzmonOrchestrate a structured writing pipeline with a source-of-truth bible and chapter consistency checks.