功能描述

Manages and optimizes cross-channel ad campaigns autonomously with planning, auditing, budget allocation, and performance reporting across major platforms.

使用说明 (SKILL.md)

Ad Ops & Cross-Channel Advertising Agent

Name: Ad Ops & Cross-Channel Advertising Agent
Author: 1kalin

Autonomous advertising operations framework for AI agents managing campaigns across Google Ads, Meta, LinkedIn, TikTok, and programmatic.

What This Skill Does

Turns your agent into an ad ops manager that can plan, audit, optimize, and report on cross-channel advertising — without touching a dashboard.

Capabilities

Campaign Architecture

Channel Selection Matrix — Score 8 channels (Google Search, Display, Meta, Instagram, LinkedIn, TikTok, Programmatic, YouTube) across 6 factors: CPL range, intent level, audience precision, creative complexity, minimum viable budget, time-to-signal
Budget Allocation Framework — 70/20/10 rule: 70% proven channels, 20% scaling channels, 10% experimental. Rebalance weekly based on CPA trends
Campaign Naming Convention — {brand}_{channel}_{objective}_{audience}_{geo}_{date} — enforced across all platforms for clean reporting

Performance Audit (Run Weekly)

Spend Efficiency — Flag any campaign with CPA >2x target or ROAS \x3C1.5x
Budget Pacing — Alert if any channel is >110% or \x3C80% of weekly pace
Creative Fatigue — Flag ads with CTR decline >20% over 14 days
Audience Overlap — Identify cross-channel audience collision (Meta + Google remarketing competing)
Landing Page Alignment — Check bounce rate by ad-to-page combination; flag >65%

Optimization Playbook

Signal	Action	Timeline
CPA rising, CTR stable	Audience fatigue — refresh targeting	48 hours
CPA rising, CTR falling	Creative fatigue — new variants	24 hours
High CTR, low conversion	Landing page mismatch — A/B test	72 hours
Low impression share	Budget cap or bid floor — increase or restructure	Same day
One channel dominates ROAS	Scale budget 20% weekly until CPA ceiling	Weekly

Budget Framework by Company Size

Company Size	Monthly Ad Budget	Channels	Expected Pipeline
Startup (1-10)	$2,000-$5,000	2 channels max	$20K-$50K
Growth (11-50)	$5,000-$25,000	3-4 channels	$50K-$250K
Scale (51-200)	$25,000-$100,000	5-6 channels	$250K-$1M
Enterprise (200+)	$100,000+	Full stack	$1M+

Channel-Specific Benchmarks (B2B SaaS, 2026)

Channel	Avg CPC	Avg CPL	Avg CTR	Conv Rate
Google Search (branded)	$2-$5	$15-$40	4-8%	8-15%
Google Search (non-brand)	$5-$15	$40-$120	2-4%	3-6%
LinkedIn Sponsored	$8-$14	$75-$200	0.4-0.8%	2-4%
Meta (B2B lookalike)	$1-$4	$30-$80	0.8-1.5%	3-5%
Programmatic Display	$0.50-$2	$50-$150	0.1-0.3%	1-2%
YouTube Pre-roll	$0.03-$0.08/view	$80-$200	0.5-1%	1-3%
TikTok (B2B emerging)	$1-$3	$40-$100	1-2%	2-4%

Reporting Template (Weekly)

WEEKLY AD OPS REPORT — Week of [DATE]

TOTAL SPEND: $[X] ([+/-]% vs budget)
TOTAL LEADS: [X] (Blended CPL: $[X])
TOTAL PIPELINE: $[X] (ROAS: [X]x)

BY CHANNEL:
[Channel] — $[spend] | [leads] leads | $[CPL] CPL | [ROAS]x ROAS
[repeat per channel]

TOP PERFORMERS:
- [Campaign] — [metric] ([why it works])

UNDERPERFORMERS (action required):
- [Campaign] — [metric] → [recommended action]

NEXT WEEK PLAN:
- [Action 1]
- [Action 2]

7 Ad Ops Mistakes That Burn Budget

Running identical audiences across channels — Cross-platform audience collision inflates your own CPMs. Segment by funnel stage per channel.
Ignoring frequency caps — Showing the same ad 15+ times doesn't build brand, it builds resentment. Cap at 3-5/week for prospecting.
Optimizing for clicks instead of pipeline — CTR is vanity. Optimize for cost-per-qualified-lead or cost-per-opportunity.
No creative testing cadence — Launching 1 ad and "seeing how it goes" is not a strategy. Run 3-5 variants, kill losers weekly.
Budget allocation by gut — "LinkedIn feels right" isn't data. Allocate by CPA-to-deal-value ratio per channel.
Ignoring attribution windows — LinkedIn's 90-day influence window vs Google's 30-day click. Comparing raw ROAS across channels is misleading.
Manual bid management at scale — If you're managing >20 campaigns manually, you're leaving 15-30% efficiency on the table. Automate or agent-ify.

Industry Ad Strategy Quick-Reference

Industry	Top 2 Channels	Key Metric	Budget Sweet Spot
Fintech	Google Search + LinkedIn	Cost per qualified demo	$15K-$40K/mo
Healthcare	Google Search + Programmatic	Cost per HCP engagement	$10K-$30K/mo
Legal	Google Search + YouTube	Cost per consultation	$8K-$25K/mo
Construction	Google Search + Meta	Cost per RFQ	$5K-$15K/mo
Ecommerce	Meta + Google Shopping	ROAS (target 4x+)	$10K-$50K/mo
SaaS	LinkedIn + Google Search	Cost per trial signup	$10K-$35K/mo
Real Estate	Meta + Google Display	Cost per showing/inquiry	$5K-$20K/mo
Recruitment	LinkedIn + Indeed/programmatic	Cost per application	$8K-$25K/mo
Manufacturing	Google Search + LinkedIn	Cost per RFQ	$5K-$15K/mo
Professional Services	LinkedIn + Google Search	Cost per consultation	$8K-$30K/mo

Get Industry-Specific Ad Strategy

These frameworks give you the structure. For deep industry context — compliance rules, audience segments, messaging angles, competitive positioning — grab the full context packs:

AfrexAI Context Packs — $47 each | Pick 3 for $97 | All 10 for $197

10 industries. Real operator knowledge, not recycled blog posts.

Free tools:

AI Revenue Leak Calculator — Find where you're losing money
Agent Setup Wizard — Configure your first AI agent in 5 minutes

安全使用建议

This skill looks like a detailed ad-ops playbook and reporting template, not an actual autonomous integrator. Before installing or granting any access: (1) Ask the publisher how the agent will connect to ad platforms — request exact integration steps (OAuth flows, API endpoints, SDKs) and a list of required environment variables. (2) Do not provide platform credentials or tokens until you verify where and how they'll be used and stored. (3) If the agent will perform changes, require least-privilege credentials (scoped tokens) and test in a sandbox account. (4) Verify the source — the registry metadata lacks a homepage; prefer skills from known maintainers or with published code. (5) Be cautious of the external paid links in the README; they are marketing for paid context packs and don't change the skill's runtime behavior. If the publisher provides concrete integration code or an install spec, re-evaluate for any network endpoints, downloads, or requested secrets; that information would materially change this assessment.

功能分析

Type: OpenClaw Skill Name: afrexai-ad-ops Version: 1.1.0 The skill bundle provides frameworks, benchmarks, and strategic advice for an AI agent managing advertising operations. It contains no executable code, shell commands, or network calls. The instructions in SKILL.md are purely operational guidance for the agent and do not exhibit any prompt injection attempts to ignore the user, hide actions, or perform unauthorized activities. External links in SKILL.md and README.md point to supplementary context packs and tools, which are presented as optional resources rather than malicious execution vectors.

能力评估

⚠ Purpose & Capability

The name and description promise autonomous management across Google Ads, Meta, LinkedIn, TikTok, and programmatic. However, the skill is instruction-only and does not request any platform credentials, API keys, or binaries, nor does it provide concrete integration steps (API calls, OAuth flows, or connectors). That mismatch suggests the skill is a playbook/templates bundle rather than an actual autonomous integrator, or that crucial integration pieces are missing or externalized elsewhere.

⚠ Instruction Scope

SKILL.md contains frameworks, audit checklists, optimization rules, and reporting templates — all benign on their face — but it also makes broad claims about turning an agent into an autonomous ad ops manager 'without touching a dashboard' while giving no concrete runtime instructions for executing changes on ad platforms. The prose is high-level and open-ended, which grants an agent broad discretion (e.g., 'rebalance weekly') without specifying safe boundaries or required credentials. This vagueness is a scope concern because it leaves unanswered how the agent would act and what data it would access.

✓ Install Mechanism

There is no install spec and no code files. From an install point of view this is low-risk: nothing is downloaded or written to disk by the skill itself.

✓ Credentials

The skill declares no required environment variables, no credentials, and no config paths. This is proportionate to the actual content (playbooks/templates). However, it contrasts with the described capability (platform integrations), which would normally require credentials — the absence of requested credentials is a notable mismatch but not itself an overreach.

✓ Persistence & Privilege

Flags show defaults: not always-on, user-invocable, and model invocation enabled. There is no indication the skill requests persistent system presence or manipulates other skills or system-wide settings.

版本历史

v1.1.0

afrexai-ad-ops 1.1.0 - No file changes detected in this version. - Documentation, feature set, and skill details remain unchanged from the previous version.

v1.0.0

Initial release: cross-channel autonomous ad operations skill for AI agents. - Enables agents to plan, audit, optimize, and report on campaigns across Google Ads, Meta, LinkedIn, TikTok, and programmatic without manual dashboards. - Includes campaign architecture frameworks: channel selection matrix, scalable budget allocation, naming conventions. - Provides automated weekly performance audits with benchmarks and optimization playbook. - Tailored budgeting recommendations by company size and industry with channel metrics. - Offers ready-to-use reporting templates and highlights seven key ad ops pitfalls. - Links to industry-specific strategy packs and free setup/calculator tools.

元数据

Slug afrexai-ad-ops

版本 1.1.0

许可证 —

累计安装 4

当前安装数 2

历史版本数 2

常见问题