worklin-brand-brain

Read and persist brand voice rules, corrections, and campaign outcomes in a stored Brand Brain profile.

1|Updated Mar 30, 2026
One-click install
npx skills add https://github.com/Logarn/Worklin-ai --skill worklin-brand-brain-logarn
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: worklin-brand-brain
Source: https://github.com/Logarn/Worklin-ai/tree/main/assistant/src/config/bundled-skills/worklin-brand-brain
Command: npx skills add https://github.com/Logarn/Worklin-ai --skill worklin-brand-brain-logarn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @vellumai/retention-domain.

What problem does it solve? Retention marketing teams lose brand consistency when voice rules, approved phrases, and campaign learnings live in scattered notes. This Skill maintains a persisted Brand Brain per brand so every generated campaign reflects approved messaging and verified past results. ## Core Features & Use Cases - Brand Profile Retrieval: Read the persisted Brand Brain bound to a conversation, or select one by brand ID, name, or website. - Explicit Corrections: Persist user-approved voice rules, taglines, CTAs, disclaimers, and forbidden claims with an audit event, only after direct user approval. - Campaign Learning Records: Store user-confirmed campaign outcomes with concrete evidence notes as dated priors, never invented results. - Research Persistence: Save completed public, read-only brand research reports (brand_research_v1) with evidence, competitor dossiers, and visual evidence validation. - Use Case: After a user confirms that a winback email outperformed a discount campaign, record that outcome with its evidence so future campaign drafts treat it as a verified learning. ## Quick Start Ask the assistant to read the Brand Brain for your brand name and then record the campaign result you just confirmed.

Frequently Asked Questions about worklin-brand-brain

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I persist a brand voice correction in Brand Brain?▼

Call brand_brain_apply_correction with the field, operation, value, and explicitly_approved set to true. Only do this after the user directly corrects or approves a durable rule; never infer preferences from a single edit or generated copy.

How do I record a campaign outcome with evidence?▼

Use brand_brain_record_campaign_outcome with the campaign type, insight, outcome (winning, mixed, or avoid), a concrete evidence note, and result_confirmed set to true. Only record outcomes the user actually supplied or confirmed.

What safety flags does a brand research report require?▼

A brand_research_v1 report must declare readOnly, publicSourcesOnly, and unsupportedClaimsExcluded as true in its safety object. Reports without these flags are rejected, and findings remain unapproved observations rather than voice rules.

Can Brand Brain store generated claims or customer quotes?▼

No. Brand Brain never stores generated claims, invented customer language, speculative outcomes, or private credentials. Research findings stay as observations or qualified inferences, not approved claims.

Why does saving a deep brand intelligence report fail?▼

Deep reports fail without the stable brand_id created during onboarding, when the report brandId mismatches, or when visual evidence items lack a research module. The report must also pass the brand intelligence contract quality evaluation.