| name | brand-audit |
| description | Runs the foundation + strategy layer of the creative-strategy framework to produce a durable brand-context bundle. Brand-intake, product-catalog, review-audit, brand-relevant-keywords, competitor-analysis, and creative-strategy-engine, run in order and saved as per-skill markdown files under /agent/brain/brand-audit/<workspace-slug>/. Every creative-deep-dive, every weekly deck, every concept brief reads this bundle as upstream context — it does not get re-run on every creative review. Refreshes once a week to catch new reviews, competitor moves, and persona shifts. Triggers on first install via missing brand-context, explicit invocation ("audit the brand", "build the brand context", "run brand-audit", "refresh the brand"), or the Monday refresh routine. |
Brand Audit
You are building the durable strategic context for a brand. Not the paid acquisition strategy — that lives in paid-strategy-audit. The brand-level layer underneath it: who this brand is, what it sells, what its customers actually say in reviews, what its competitors are doing, and the pain × persona × messaging-angle matrix every concept brief and creative diagnosis draws from.
The output isn't a report. It's six durable markdown files that every downstream skill reads as context. It runs once on setup and refreshes weekly. It is not re-run every time someone wants to review a creative — that would be a waste.
Core principle
This is the brand layer, not the paid layer. It sits next to /agent/brain/paid-strategy/ because it provides the strategic foundation those briefs draw from, but it is not channel-specific. One brand, one bundle. If the brand operates on Meta and TikTok and Pinterest, the brand-audit still runs once. Channel-specific paid strategy lives in paid-strategy-audit.
Durable artifacts, not narrative outputs. Every file in the bundle is something a downstream skill reads programmatically. Structure matters. Section headers must match what the skills expect.
Refresh weekly, not on every read. The Monday morning refresh routine catches new reviews, new competitors, persona shifts. The deep-dive doesn't trigger a re-audit — it just reads the latest bundle.
What this skill produces
Per workspace, in /agent/brain/brand-audit/<workspace-slug>/:
brand-context.md — identity, value prop, audience, positioning. Output of brand-intake.product-catalog.md — product list, benefits, claims, price points. Output of product-catalog.review-audit.md — VOC across pain, desire, objection, comparison, and use-case buckets. Output of review-audit.keywords.md — search and language landscape. Output of brand-relevant-keywords.competitor-analysis.md — who they actually compete with, what those competitors are doing in ads, content gaps. Output of competitor-analysis.strategy.md — the pain × persona × messaging-angle matrix mapped across the 5 awareness stages. Output of creative-strategy-engine. This is the file the deep-dive's Strategic Layer section reads on every single creative.
Plus:
_history/<archived-iso>/ — full prior bundle preserved on every weekly refresh._config/<workspace-slug>.json — refresh schedule, ping channels, input sources (brand site URL, competitor list, review sources).- The
runneth-apps:brand-audit:read-before-creative v1 block appended to /agent/user.md (idempotent, sentinel-guarded).
Execution
Step 0 — Resolve scope
- Workspace = Motion context default unless the user named another.
- Check
/agent/brain/brand-audit/<workspace-slug>/brand-context.md. If present, this is a refresh, not a first build — keep the existing files in memory for the diff at Step 7.
- Read the workspace's input sources from
_config/<workspace-slug>.json if present, otherwise route to the setup skill first.
Step 1 — Run brand-intake
Invoke the brand-intake skill against the configured brand site URL. The skill produces a structured brand-context output covering: brand identity, value proposition, primary audience definitions, positioning vs the category, and the brand's own voice and tone signals.
Save the output to /agent/brain/brand-audit/<workspace-slug>/brand-context.md.
Step 2 — Run product-catalog
Invoke the product-catalog skill against the brand site and any uploaded product docs. Produces the full product list with benefits, claims, price points, and any constraint a downstream creative skill needs to know about.
Save to /agent/brain/brand-audit/<workspace-slug>/product-catalog.md.
Step 3 — Run review-audit
Invoke the review-audit skill against the configured review sources (Trustpilot, Amazon reviews, internal NPS exports, App Store, G2, anything the team has). The skill mines reviews and categorizes voice-of-customer signal across five buckets: pain points, desires, objections, comparisons, and use cases. Each bucket gets representative verbatim quotes.
Save to /agent/brain/brand-audit/<workspace-slug>/review-audit.md.
If no review sources are configured, skip this step and flag it as a gap in the brand-context summary. The Monday refresh will pick it up if sources get added later.
Step 4 — Run brand-relevant-keywords
Invoke the brand-relevant-keywords skill against the brand + product catalog + competitor list. Produces the search-landscape and language map: high-intent keywords, audience-language phrases, category terms, branded terms, and the gap between how the brand talks and how the customer searches.
Save to /agent/brain/brand-audit/<workspace-slug>/keywords.md.
Step 5 — Run competitor-analysis
Invoke the competitor-analysis skill against the configured competitor list (or auto-detected competitors if none configured). Produces: who the actual competitors are by category position, what each is currently doing in ads (using motion inspo-creatives for active competitor ads), where the content/messaging gaps are, and the brand's strategic moat-vs-vulnerability read.
Save to /agent/brain/brand-audit/<workspace-slug>/competitor-analysis.md.
Step 6 — Run creative-strategy-engine
Invoke the creative-strategy-engine skill with brand-context, product-catalog, review-audit, keywords, and competitor-analysis as inputs. Produces the pain × persona × messaging-angle matrix mapped across the 5 awareness stages (Unaware → Problem-Aware → Solution-Aware → Product-Aware → Most-Aware).
This is the single most-referenced file in the downstream creative work — creative-deep-dive's Strategic Layer section reads it on every per-creative run, and weekly-performance-deck's recommendations slide reads it to make sure proposed iterations fit a real persona × pain anchor.
Save to /agent/brain/brand-audit/<workspace-slug>/strategy.md.
Step 7 — Refresh diff (on re-runs only)
If this is a refresh, compare each new file against its prior version. Surface a short summary of what changed:
- New customer pain points appearing in review-audit
- New competitors entering the analysis or existing competitors making material moves
- New persona or messaging angle introduced in the strategy matrix
- Material shifts in product-catalog (new products, price changes, retired claims)
If anything material changed, post to the configured ping channels (same channels as the paid-strategy-audit drift):
Weekly brand-audit refresh for <workspace>. . . React ✅ to adopt, ❌ to roll back to prior version, ❓ to say you'll review.
Archive the prior bundle to _history/<archived-iso>/ regardless of whether the new one gets adopted. On approval (or 48h silence + only-additive changes), the new bundle replaces the prior. On rejection, restore from history.
Step 8 — Install user.md snippet (first run only)
Idempotent via sentinel. If /agent/user.md does not contain runneth-apps:brand-audit:read-before-creative v1, append:
<!-- runneth-apps:brand-audit:read-before-creative v1 -->
### Creative work — read the brand audit first
Before any creative-strategy work (deep dives, hook writing, concept briefs, ad teardowns, brand-relevant copy, weekly deck recommendations) for a workspace with a brand-audit, read the relevant files under `/agent/brain/brand-audit/<workspace-slug>/` first:
- `strategy.md` for the pain × persona × messaging-angle matrix and awareness-stage map (almost always needed)
- `brand-context.md` for identity, value prop, voice and tone
- `product-catalog.md` for benefits, claims, price points, constraints
- `review-audit.md` for voice-of-customer signal across pain, desire, objection, comparison, use-case buckets
- `keywords.md` for audience-language vs how the brand talks
- `competitor-analysis.md` for what competitors are running and where the gaps are
Read narrowly — pull only the file(s) the current task needs. The bundle is durable; the brand-audit refresh on Monday morning keeps it current. Do not re-run brand-audit on every creative review — read the existing bundle.
<!-- /runneth-apps:brand-audit:read-before-creative -->
Step 9 — Schedule the Monday refresh (first run only)
reminder add --title "Weekly brand-audit refresh" \
--description "Re-run brand-audit for workspace <workspace-id>. Refresh review-audit and competitor-analysis (most volatile), re-derive strategy matrix on any material shift, ping configured channels for confirmation on material changes." \
--schedule "every monday at 08:00 <timezone>"
Monday morning is intentional — the refresh lands before the Friday deck, so the deck can reference the latest brand context.
Step 10 — Update INDEX
Append entry to /agent/INDEX.md:
- path: /agent/brain/brand-audit/<workspace-slug>/ | aliases: brand audit, brand context, brand-strategy bundle, customer-voice, persona matrix, competitor read | note: Durable brand-strategy bundle for <workspace> produced by brand-audit. brand-context.md (identity + voice), product-catalog.md, review-audit.md (VOC), keywords.md, competitor-analysis.md, strategy.md (pain × persona × angle × awareness matrix — most-referenced). Refreshes Monday mornings. Read by creative-deep-dive and weekly-performance-deck. | created: <date> | updated: <date>
Standalone skill invocation
Every one of the six skills inside brand-audit is also independently callable. Users can run any of them in isolation:
brand-intake — quick brand-context rebuild from a new URLproduct-catalog — refresh product list after a launchreview-audit — VOC pass on a new review corpusbrand-relevant-keywords — keyword landscape for a campaigncompetitor-analysis — refresh competitor readcreative-strategy-engine — regenerate the persona × angle matrix
When invoked standalone, each skill writes its output to the same per-workspace path under /agent/brain/brand-audit/<workspace-slug>/. The full orchestrator just runs them in order.
Anti-patterns
- Don't re-run brand-audit on every creative deep dive. Read the existing bundle. Refresh weekly.
- Don't store the brand-audit output inside
/agent/brain/paid-strategy/. Brand is brand-level, not channel-level. It sits next to the paid-strategy library, not inside it.
- Don't skip review-audit silently when no sources are configured. Flag the gap; the bundle is incomplete without VOC.
- Don't merge the strategy matrix into the brand-context file. Downstream skills expect six separate files at known paths. Don't break the contract.
- Don't ping a DM with the weekly refresh diff. Channels only, same as paid-strategy-audit.
What this bundle enables
Any future agent reading the brand-audit bundle can answer:
- What does this brand actually sell, to whom, against whom?
- What do customers actually say about it (verbatim)?
- Which pain × persona combinations have validated messaging angles for them?
- What's the right awareness stage for a given concept?
- Where are competitors gapping the brand right now?
- Which audience-language words and phrases are reusable in copy?
creative-deep-dive's Strategic Layer section is grounded in strategy.md. weekly-performance-deck's recommendation slides cross-check proposed iterations against strategy.md's persona × pain matrix to make sure the team isn't shipping creative for a persona that isn't validated. Any future hook-writing, concept brief, or landing-page skill reads the same bundle.
