Blog Brand Context
Create durable brand positioning and voice guidance for every blog workflow.
The skill requires an interactive interview and confirmation before overwriting existing files, and it collects brand, audience, competitor, and disclosure context. However, it writes BRAND.md and VOICE.md to the project root and causes them to be auto-loaded into downstream prompts without documenting sensitive-data handling, access scope, backup, or rollback in this skill. Security assurances mainly depend on the external orchestrator and helper scripts, so points are deducted.
Commands, interview steps, output templates, and abnormal cases are largely consistent, including handling missing personas, unclear project roots, and existing files. However, the skill depends on the blog orchestrator, blog-persona, and their conventions, and provides no independently reproducible execution path, standardized failure output, or validation rules. The static-review ceiling also limits the score.
Invocation phrases, three commands, and intended use are reasonably clear for initialization, inspection, and updating. Boundaries for non-fit scenarios, Chinese-language interviewing, localization, alternate project layouts, and non-Claude-Code environments are not specified, so points are deducted.
Frontmatter, MIT licensing, argument hint, command table, templates, relationship notes, and error handling are useful and readable. The skill itself lacks a version, changelog, named maintenance responsibility, and explicit update path, and it relies on unspecified details from other skills and the orchestrator.
The generated BRAND.md and VOICE.md structures are directly usable for persistent editorial rules and downstream blog-skill consumption. Actual file generation, automatic loading, and persona consistency depend on other components, and no skill-specific execution evidence is provided, so the score remains moderate.
The supplied repository context includes CI and test claims, but the shown tests primarily cover other scripts and the delivery contract, not blog-brand interviewing, overwrite confirmation, template generation, or automatic loading. Evidence is therefore only indirectly auditable.
- BRAND.md and VOICE.md enter downstream system prompts; do not store secrets, private data, or unreviewed instruction-like content in them.
- Initialization and update can overwrite project-root files, but no backup or recovery workflow is defined; version-control and review the diff first.
- Core behavior depends on the blog orchestrator and blog-persona; standalone installation or missing components may have unclear boundaries and failure behavior.
- The supplied material does not establish Chinese interview support or reachability in mainland-China network environments.
What does this skill do, and when should you use it?
Blog Brand is a focused skill in the claude-blog suite for establishing reusable brand and voice context. It runs an interactive interview and writes BRAND.md and VOICE.md at the project root, covering audience, positioning, editorial rules, topic boundaries, and style preferences. When those files exist, the blog orchestrator loads them before writing or reviewing content. It is best suited to creators and teams that need consistent editorial decisions across sessions and blog commands.
The skill creates or updates BRAND.md and VOICE.md in the project root. Its five-step interview collects audience details, positioning, editorial rules, topic scope, and voice preferences; it supports init, show, and update operations. If a blog-persona JSON file exists, voice answers are prefilled from it. show checks whether both files exist and displays a summary; update uses current values as defaults before overwriting the files.
- A marketing lead needs consistent audience definitions, positioning, and competitive differentiation across a B2B blog.
- A content team wants durable do-not-use phrases, disclosure requirements, and editorial rules for every downstream writing task.
- A solo creator wants the same pronoun stance, headline patterns, sentence ceiling, and paragraph ceiling across multiple sessions.
- A team using blog-persona wants readable brand context that can be loaded by all blog sub-skills.
What are this skill's strengths and limitations?
- Centralizes audience, positioning, editorial rules, and voice guidance as project-level context.
- Provides explicit init, show, and update workflows for setup and maintenance.
- The blog orchestrator can inject the files into writing, rewriting, briefing, outlining, calendar, and strategy workflows.
- Complements blog-persona by supplying readable cross-skill context while persona JSON remains the structured voice source.
- It primarily manages two Markdown context files; it is not a complete blog-writing or SEO-audit system.
- Automatic loading depends on the claude-blog orchestrator, so standalone usefulness is limited.
- Users must provide substantive interview answers; the skill refuses to write skeletal documents when answers are too minimal.
- The supplied material documents no installation or execution path independent of Claude Code.
How do you install this skill?
The skill is distributed with the public AgriciDaniel/claude-blog plugin collection. With Claude Code 1.0.33+, run /plugin marketplace add AgriciDaniel/claude-blog, then /plugin install claude-blog@agricidaniel-claude-blog. The skill source is skills/blog-brand/SKILL.md.
How do you use this skill?
In Claude Code, run /blog brand init to start the five-step interview and create BRAND.md and VOICE.md. Run /blog brand show to inspect their status and summary. Run /blog brand update to revisit the interview with existing values prefilled. If the project root is unclear, the skill asks where to write the files.
How does this skill compare with similar options?
Compared with giving Claude or ChatGPT brand instructions in each prompt, Blog Brand persists the guidance in project files for reuse across sessions. Compared with blog-persona, it focuses on audience, positioning, taboo phrases, disclosures, editorial boundaries, and readable voice context rather than structured persona data alone.