blog-persona

Blog Persona is a voice-and-tone management skill within the claude-blog toolkit. It solves the problem of inconsistent brand voice across blog content by capturing a brand's writing style as a structured, reusable persona. Built on the NNGroup 4-dimension tone framework (Funny-Serious, Formal-Casual, Respectful-Irreverent, Enthusiastic-Matter-of-fact) and the CMI Brand Voice Chart, a persona encodes readability targets, sentence-length distribution, vocabulary tier, contraction frequency, do/don't lists, and a preferred summary-box label so that every post reads as if written by the same author.

The skill exposes commands to create, list, use, and show personas. Creation runs a six-step interactive interview covering brand basics, tone-dimension sliders, writing rules, do's and don'ts, summary-label preference, and optional voice-sample URLs. Personas are saved as JSON files (kebab-case, with path-traversal protections) under the skill's references directory, following a documented schema. Once a persona is active, blog-write and blog-rewrite load it to inject tone and style rules into generation and to validate output against the persona's constraints. Notably, the optional voice-sample step includes explicit, security-conscious fetching rules: allow only http/https, reject javascript/data/file URLs, block loopback and private IP ranges to prevent SSRF, validate redirects, cap size and timeout, and treat fetched page text as untrusted data rather than instructions.

The target users are content teams, marketers, and solo bloggers who publish frequently and need to enforce a coherent voice. Use cases include codifying a company's B2B SaaS voice, maintaining separate personas for different brands or audiences, and ensuring AI-generated drafts match an established editorial style.