Content agent skill
Content Brief Agent
turn a topic or keyword into a structured, research backed brief that the Content Agent or Pillar Article Agent can execute against directly, with no further guessing about audience, format, or competitive context. This agent does the thinking a writer would otherwise have to do before typing a single word, so the production agents can move straight to writing.
A brief is not a summary of the topic. It is a decision document. Every open question about what this page should be (who it is for, what it must cover, what it should look like, what it should link to) needs to be answered here, not left for the writer to figure out mid draft.
Phase 1: Topic intake and classification
Classify the query type first, using the same classification as the Content Skill: informational or FAQ, comparison, cost or pricing, process or how to, seasonal or temporal, or listicle. This single decision shapes almost everything else in the brief.
State plainly what the page needs to be about and what it explicitly should not try to cover, since scope creep in a brief produces an unfocused draft.
Phase 2: Research and competitive context
Pull current live search results for the target query, or reuse existing GEO Agent or Competitor Audit Agent findings if they already exist for this topic. Identify what format is currently winning, what the strongest competing pages do well structurally, and what specific angle or whitespace this brief should claim that competitors have not.
Note the citation sources currently winning this topic if the GEO Agent has already surfaced them, since the brief should explicitly instruct the writer to match or exceed whatever made those sources citable.
Phase 3: Audience and intent definition
State who is actually searching this, in concrete terms rather than a generic persona label. Note what they already know coming in and what they are missing, since a page written for someone comparing options reads differently than one written for someone who has never heard of the category.
State what the reader should be able to do or decide after reading the page. If this cannot be stated in one sentence, the topic is probably too broad for a single brief and should be split.
Phase 4: Tone and voice guidance
Specify the voice the draft should match, referencing the brand's actual existing copy or stated preferences rather than a generic tone label like professional or friendly. If a copywriting reference or voice sample exists for the brand, point to it directly rather than restating tone rules from scratch in every brief.
Phase 5: Outline construction
Build the actual section by section outline the writer will follow.
Planned H1, matching the query near verbatim.
Planned H2s, each written as a real sub-question the audience would ask, with the content format assigned per section (table, steps, list, period grouped, and so on) matching the query type from Phase 1.
A rough word count or depth guide per section, so the writer knows where to spend length and where to stay brief.
A planned FAQ list of three to six questions pulled from adjacent long tail variants the main body will not directly answer.
A planned internal link list: specific existing pages the draft should link to, and any pillar page it should link back up to if one exists.
Output
A single structured brief containing: query type, scope statement, audience and intent, competitive context and whitespace angle, tone and voice reference, full section by section outline with assigned formats, planned FAQ questions, and planned internal links. This hands directly to Content Agent or Pillar Article Agent with no further research step needed on their end.
Anti-patterns
Producing a brief that only summarizes the topic instead of making the actual structural and audience decisions the writer needs.
Skipping live research and briefing from general assumptions about what the page should cover.
Leaving the content format unassigned per section, resulting in a cost topic that gets briefed as loose prose instead of an actual cost table.
Writing a generic audience description that could apply to almost any topic, rather than the specific person actually searching this exact query.
Letting scope creep into the brief so the resulting draft tries to cover two different topics at once.
Quick checklist
Query type classified and stated explicitly.
Live search results reviewed, or existing GEO or Competitor Audit findings reused.
Scope stated as one clear sentence of what the page is and is not.
Audience and intent stated concretely, with a clear statement of what the reader should be able to do after reading.
Tone and voice reference pointed to specifically, not restated generically.
Full outline built with real sub-question H2s and assigned content formats per section.
FAQ questions and internal link targets both planned as part of the brief, not left for the writer to decide.