菜单
Analyze Meta ad account data — read what's spending, what's converting, what's fatiguing, and where the audience is being reached. Use when the user wants an account audit, a single-ad diagnosis, a placement read, a demographic breakdown, or a fatigue check.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
How this brain saves, syncs, and clones itself on GitHub — the required procedure for EVERY git push, pull, fetch, clone, commit, or gh command in a Parker Brain repo. Covers the credentials (the setup_parker_brain tool, never the user's own login, never gh), submodules, merge conflicts, and the rare self-hosted exception. Use whenever saving work to GitHub, syncing or updating the repo, cloning the brain, fixing a failed push or auth error, or when asked to save, back up, or share the brain.
The front door to building a brand's brain. Run this when standing up a new brand — all it needs is the Parker MCP connected. It welcomes the person, provisions the brand's own repo through the setup_parker_brain tool (Parker MCP), clones that repo and works inside it, connects and tests the data, asks the short brand intake (the things Parker genuinely can't observe), runs the full build in dependency order, saves it to GitHub, and hands off to get-started. Follows prompts/onboarding-runner.md as the detailed method. Use when someone says set up a brand, build the brain, onboard a new brand, or get started building. Also the right entry when the brand's brain already exists — it detects a built brain and retrieves it, handing off to get-started instead of rebuilding.
The weekly check that keeps this brain current with the parker-brain standard — without overriding anything the team chose. Compares the pinned parker-system/ release against the newest factory tag and offers the bump as one short question; on a yes the migrations run and a script re-syncs the copied skills deterministically. What the team declines stays declined; what they modified stays theirs. Use weekly as a scheduled routine, or when asked to update the brain, check for updates, sync with the standard, or "what's new in Parker".
One-time setup for a fresh brand brain instance. Registers the standing routines (refresh-context, dream, harvest+evaluate ideas, research-loops, update-brain, self-improve) as scheduled cloud agents so they run on cadence without being asked. The job definitions already travel with this repo; this skill arms the schedules, which are per-account and can't be committed. Run once after cloning the brain into a new Claude Code cloud instance, or when asked to set up / register / schedule the routines.
Generate AI ad assets — Veo 3 video prompts and AI image static prompts — with brand context, compliance, and forensic direction. The output is prompt text the user pastes into the AI tool, not the final asset.
Write headlines for ads — static, video text overlays, paid social, landing page hero. Picks the right structural pattern for the brand type (lifestyle vs problem/solution), the awareness stage, and the customer language available.
| name | ad-account-analysis |
| description | Analyze Meta ad account data — read what's spending, what's converting, what's fatiguing, and where the audience is being reached. Use when the user wants an account audit, a single-ad diagnosis, a placement read, a demographic breakdown, or a fatigue check. |
| triggers | ["analyze my ad account","what's working in the account","what's the top spender","is this ad fatiguing","is this ad ready to scale","what does the data tell me","read my Meta account","audit this ad","which ads should I turn off","where is my spend going"] |
Read what the Meta data is actually saying about the account, an ad, or a slice of the audience. Account data tells two intertwined stories — whether ads are scaling profitably, and how the ads are behaving in the Meta ecosystem in terms of attention and audience reach. The skill reads both stories, decides which one the user needs, and delivers a diagnosis with specific actions.
This skill does not write ads, generate iterations, or build briefs. If the user wants iteration recommendations off the analysis, route to the iterations skill. If they want a script or static built from a diagnosis, route to the scriptwriting or static-ads skill.
This skill reads the account through the canonical methods, not from general knowledge. Before analyzing, load what parker-system/creative-strategy-context/expertise-routing.md names for an own-account read: ad-account-analysis.md, the canonical reading method; killer-performance-ads.md, the bar a winner is graded against; and andromeda-v2.md, so delivery patterns read as Meta auction mechanics rather than mysteries. The tools that pull the data, and what each returns, are inventoried in parker-system/system/parker-tools.md. When the read compares how cohorts of ads acquire, the cohort-efficiency discipline in creative-strategy-fundamentals.md applies: spend-weight the comparison and classify off the account's own creative tags, never best-ad-vs-best-ad. The diagnosis is performed through these methods, in their vocabulary — a read that never speaks their named concepts skipped them.
Two pieces of background context every account analysis depends on. Both are defined in full in ad-account-analysis.md — that doc is the source of truth, including the worked breakdown-effect example and the full attribution-window list. What follows is only the must-surface summary so the read never forgets them. Parker assumes both, but if the user's question hinges on either, surface them explicitly.
Meta's delivery system shifts budget toward whichever ad set, placement, or ad is predicted to drive the most efficient outcomes over the campaign duration, not over the moment Parker happens to look at the data. That means a placement or ad that currently shows a higher CPA may be the right place for budget because Meta predicted its costs would rise more slowly than the alternative. Reporting in Ads Manager can therefore look like Meta made the wrong call when it actually did not. Always evaluate at the level appropriate to the campaign structure — at the campaign level for Advantage+ campaign budget, at the ad set level for automatic placements, at the ad set level when multiple ads sit in one ad set.
Different attribution settings count conversions differently. Click-through credits actions within one or seven days of a click. View-through credits actions within one day of an impression. Engaged view credits actions after ten seconds of video play. Standard attribution combines a window with a behavior. Incremental attribution optimizes for conversions the model predicts the ad caused. The brand's chosen attribution setting changes which numbers tell the truth. If the user has changed attribution settings between two periods, account-level comparisons mean different things.
Identify what the user is actually asking. Run the strategy doc. Are they asking for a full account audit, a single-ad diagnosis, a placement read, a demographic breakdown, or a fatigue check? Different questions trigger different processes.
Pull the data needed for the chosen process. Each process has its own required-input list. Pull only what is needed.
Apply both lenses — profitability and user behavior. Every analysis must consider both whether the ad is making money (profitability metrics) and how the ad is being received in the ecosystem (engagement metrics). Single-lens analysis misses the picture.
Diagnose, do not just report. The output is not a metrics dump. The output is the read — what is true about the account or ad, what is the explanation, what should the team do next.
Format the output per the structure below.
Two to four sentences. Plain language. State what the data is saying and what the user should do about it. No jargon. If the answer is "this ad is fatiguing and should be iterated within two weeks," say that.
The supporting evidence. Specific numbers compared against account benchmarks. Both lenses present — profitability and user behavior. Tied to the actual question the user asked.
Action items, prioritized. Each item names what to do and why, tied to the diagnosis. If the next step is running a different skill (iterations, scriptwriting), name the skill.
What brand-specific KPIs were used to define "winning," what compliance or operational constraints shaped the read, and why the diagnosis fits the brand's current situation.
