Skip to content

gooseworks-ai/goose-skills / skills/marketing/voice-card

voice-card

Verified

Builds or updates a voice card from a company blog ("extract our blog voice") or how an account or person actually writes, measured from about 8 or more of their real posts (length, casing, emoji, openings, pet phrases, opinions with their source). When there are too few posts, it runs a short interview instead. It ends with a test that checks a draft against the card. The brand's rules always beat the card. Use alone when you already have the account's posts or the person on hand; otherwise run the playbook for the job (e.g. run-organic-social, write-outreach, email-lifecycle).

FileKindSize
SKILL.mdskill md6.9 KB
references/blog-voice.mdmarkdown4.7 KB
skill.meta.jsonconfig655 B
tests/examples/2026-10-22/blog-voice.mdmarkdown7.1 KB
tests/fixtures/company-blog.mdmarkdown6.8 KB
tests/test-by-hand.mdmarkdown2.4 KB

tests/test-by-hand.md

View on GitHub
# Blog voice acceptance — 22 October 2026

This tests a sourced company-blog guide within the existing voice-card step. All inputs are original fictional fixtures. It verifies the local text/measurement path; it does not claim owner confirmation or a live brand save.

## Fixture run

Input: `fixtures/company-blog.md`. Ask: “Extract our company blog voice from these five articles.” Use the existing confirmed brand rules in the fixture; owner has not confirmed a guide.

Actual output: `examples/2026-10-22/blog-voice.md` (1,072 words). Body-word counts exclude headings/metadata and include CTA; hyphenated words count once. Measured with text tools: 898 body words, five articles (177–182 each), 85 sentences (median 11 words, range 4–18), 25 paragraphs including five short CTAs; three subheadings/article. Fresh sample: 44 words, sentence lengths 9/8/7/10/10, within observed ranges. Every source link to a local fixture resolves.

Inspected source/author/date receipt; five quoted rules; article structure, CTA and terminology; three generic-to-voice rewrites; brand-rule precedence; no invented performance data; provisional status from short corpus and absent confirmation. The line-up is an agent check, not an independent human blind test. The short fixture is not presented as an ideal real article length. No company-kit voice mutation or social-card overwrite occurs.

## Repeat and failure cases

- Supply three excerpts rather than full articles: analyse what exists, provisional; unobserved article structure stays unknown.
- Give a guest post in a different voice: record the separate mode; do not average it into company voice.
- Existing confirmed rule clashes: retain it and ask for owner choice, not silent replacement.
- One sample makes an unapproved result claim: treat it as style evidence only and select a safe example.
- Owner confirms: save this blog handle as confirmed only when the brand precedence permits it; preserve the older current rules unless their change was approved.
- No body-readable fetch: incomplete/provisional guide plus exact missing read; no invented article corpus.

## Remaining live checks

The fixture run made no fetch, provider, brand-save or external message. A live host run must check public article-body retrieval, persistence/versioning and fresh-session retrieval; owner voice confirmation remains a human decision. Skill validation passed (351 skills at that point); no owned-file installed-layout link errors.
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 323kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 323kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 323kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall