Maintain a reusable brand-rule skill

Why use this

1. A design team may find brand guidance spread across documents, tokens and component libraries that do not change together. Reusing an old instruction can create consistent-looking work that no longer reflects the brand owner's intended system.

2. This job could assemble a reviewable rule set and make proposed updates explicit. Separating firm rules, contextual examples and unresolved conflicts helps the brand owner decide what should guide future work before the skill is treated as authoritative.

Why an agent / simpler alternative

Useful judgment

Reconcile prose guidance with tokens and components, distinguish exceptions and explain which rule changes need brand-owner judgment.

Simpler alternative

Use a concise brand checklist when guidance is stable and centralized.

Unsuitable when

The workflow would overwrite approved rules, expose private design assets or infer policy from isolated examples.

Proposed inputs

1. Name

   Approved brand sources

   Owner

   Brand owner

   Minimum scope

   Relevant documents, tokens and library excerpts

   Quality check

   Check versions, ownership and permitted use

2. Name

   Rule governance brief

   Owner

   Brand owner

   Minimum scope

   Priority among sources, protected rules and review scope

   Quality check

   Identify examples versus mandatory standards

Access and setup — proposed or unknown

Proposed minimum

1. Read supplied excerpts and draft a separate rule candidate/change summary; no live library edits or scheduled rebuild.

Compatibility unknowns

1. Current design connectors, skill format, source versioning and runtime compatibility are unknown.

External actions requiring approval

1. Access expansion, external asset uploads, shared-skill replacement and recurring rebuilds require authorization.

Expected deliverable and review checks

Proposed artifact

Brand-rule candidate with source pointers, exceptions, conflicts and proposed update log.

Critical checks

1. Preserve explicit rules; trace tokens to versions; distinguish examples from requirements; require brand review before hardening.

Missing output evidence

This overlay adds no new output inspection, execution or validation; consult the referenced canonical source outcome limits.

Failure signals and proposed responses

1. Signal

   A showcase component contradicts written guidance

   Why it matters

   A single example may be mistaken for policy

   Proposed response

   Surface the conflict and request the owner's precedence decision.

2. Signal

   A renamed token is treated as a new color

   Why it matters

   The guide invents a visual change

   Proposed response

   Compare actual values and record naming changes separately.

Effort and cost — unknown

Setup work

1. Define source precedence and rule categories.

Ongoing review

1. Brand owner reviews updates and removes obsolete guidance deliberately.

Unknowns

1. Source freshness, token semantics, maintenance effort and service costs are unknown.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

Invent a brand paragraph, token list and component example with one conflicting spacing value and a renamed unchanged color; draft an update summary.

Acceptance checks

1. Flag the real conflict; distinguish rename from value change; preserve protected rules; save no persistent skill.

Stop conditions

1. Stop if source precedence or permission to replace shared guidance is unresolved.

Limits

1. A reusable brand-rule candidate cannot establish universal design quality or replace context-sensitive brand judgment.