LinkedIn Ads MCP in Claude: The BlueAlpha Marketing Plugin

The BlueAlpha LinkedIn skills put a senior LinkedIn ad operator inside Claude. Ten skills connect directly to your LinkedIn ad accounts and handle the jobs that turn an account into decisions: structural health audits, demographic targeting analysis, count-blocked campaign detection, creative fatigue scoring, auction overlap between your own campaigns, bid strategy audits, frequency saturation risk, Lead Gen Form health, weekly performance digests, and a unified quarterly audit that composes all nine into one report. Each skill follows the same loop: Analyze, Decide, Plan, Act, Prove This page is the index. Pick the skill that matches your job-to-be-done and jump to the prompt that runs it.

The video walkthrough was recorded in May 2026. Account metrics (spend totals, campaign counts, demographic findings) might differ between the video and the written examples. Both reflect real data from the BlueAlpha account at their respective run dates.

If you've spent six figures on LinkedIn ads for a B2B product, you've felt the gap. You have an account with dozens of campaigns, hundreds of creatives, and a long list of audiences. You have a CFO asking whether the spend is working. You have a planning meeting on Monday where someone has to say what to cut, what to scale, and what to refresh, and LinkedIn Campaign Manager shows you twelve dashboards but no decisions. Because a dashboard is not a decision.

The BlueAlpha LinkedIn skills close that gap. Each one is an agent that turns one slice of your LinkedIn account into a real call: trust the campaign or pause it, scale the audience or exclude it, refresh the creative or double down on it, fix the structure or fix the bidding first. The skills don't replace your media buyer. They extend the operator's reach so every analysis happens in minutes instead of half-days, and so a single Head of Growth can run the loop that used to require a dedicated paid social team.

All 10 LinkedIn Skills in the BlueAlpha Suite

Read what's happening:

• linkedin-auto-optimize: The account-level pre-flight check. Surfaces structural blockers (billing holds, audience count holds, no-creative campaigns), under-pacing campaigns, and dormant ACTIVE campaigns. Lead any LinkedIn conversation with this; never optimize creative on an account that's structurally broken.
  
  

• linkedin-performance-digest: The Monday-morning LinkedIn read. Spend, impressions, LP clicks, leads, conversions period-over-period with named drivers and a what-to-watch list.
  
  

• linkedin-demographic-deep-dive: The targeting-vs-actual report. Who you're paying to reach vs who's actually clicking, broken down by seniority, job function, company size, and industry.

Decide what to do with your audiences and bids:

• linkedin-audience-health-check: Matched-audience risk surface. Surfaces count-blocked campaigns, single-point-of-failure audiences, audiences shared across active campaigns (auction overlap risk), and compound-risk audiences that are both.
  
  

• linkedin-targeting-overlap-finder: Cross-campaign auction overlap. Computes per-facet Jaccard similarity between every pair of active campaigns and flags pairs above 60% overlap with a daily-budget estimate of the wasted auction competition.
  
  

• linkedin-frequency-saturation-report: Cap configuration audit. Identifies matched-audience campaigns running without frequency caps (the highest saturation risk), caps set too tight, and caps set too loose.
  
  

• linkedin-bid-strategy-audit: Per-campaign bid configuration sanity check. Flags cost-type mismatches, suspicious unit costs, missing optimization targets, and the ROUND_ROBIN-vs-OPTIMIZED creative selection trap.

Decide what to do with your creatives:

• linkedin-creative-fatigue-watchdog: Per-creative CTR/engagement decay scoring using objective-appropriate metrics. Identifies fatigued creatives, weak-from-launch creatives, high-spend low-engagement waste, and top performers.
  
  

• linkedin-lead-form-quality-auditor: Lead Gen Form surface audit. Surfaces creatives spending without producing leads, declining-volume forms, dormant Lead Gen creatives, cost-per-lead leaderboard, and orphan forms.

Run the whole audit at once:

• linkedin-full-monty: Runs all nine sub-skills against shared data and composes the findings into a single comprehensive audit. Cross-skill prioritization: a count-held audience that's also shared in auction overlap gets ranked above either single-skill finding.

Analyze > Decide > Plan > Act > Prove

Every LinkedIn skill follows the same loop, the same one the MMM skills, the Google Ads skills, and the TikTok skills follow:

1. Analyze. Every morning, agents surface what needs attention: anomalies in campaign pacing, creative fatigue, budget drift, competitive moves, and targeting leakage. You don't go digging through Campaign Manager; the agent brings the relevant slice to you.

2. Decide. The relevant agent produces a recommendation with the signal that backs it. Every recommendation includes the dollar amount at stake, the underlying metric, and the rationale. You see the data, the logic, and the dollar impact before you act.

3. Plan. The Planning Agent translates every decision into the forward roadmap. What's running this week, what's being tested next, what budget moves are queued. The plan is the living artifact your team and your leadership actually look at.

4. Act. The agent produces a deployment-ready output: a list of audiences to expand, a creative refresh queue, a cap-policy table you can apply directly, a campaign-merge recommendation with cross-exclusions ready. Not a slide in a deck; a spec you can ship the same day.

5. Prove. After changes ship, the loop closes with a measurement check: did the move produce the predicted lift? Period-over-period reads, geo holdout results, and MMM refits feed back into the next cycle so every subsequent recommendation is sharper than the last.

A single Head of Growth with the BlueAlpha LinkedIn skills connected manages what used to require a dedicated LinkedIn media buyer plus a creative analyst plus the eternal back-and-forth with the agency on "what's the bid type supposed to be for Lead Gen again?"

Who the LinkedIn Skills Are For

• You're a Head of Growth running LinkedIn alongside Google Ads and Meta. You don't have time to learn LinkedIn's quirks: matched audience count holds, ROUND_ROBIN vs OPTIMIZED, dynamic vs matched segments. You need something to tell you what to do.

• You're a B2B CMO and you have a CFO asking whether the $30K/month you spend on LinkedIn is producing pipeline. You need an answer with dollars, not a screenshot of an impression chart.

• You're a paid media manager who knows your account is leaking money but the LinkedIn UI surfaces twelve KPIs and zero recommendations.

• You're an agency operator running LinkedIn for 5-20 clients and the manual audit work has become the bottleneck on growing your book.

Every skill is designed around the same principle: the human stays in the loop, Claude does the heavy lifting.

BlueAlpha delivers the suite through forward-deployed engineering. A senior growth partner embeds with your team, builds the measurement and execution stack directly, and trains your team to run the agents independently. You keep the seat. The agents eliminate the hours. The knowledge base stays yours.

How to Pick the Right Skill

How Does the BlueAlpha Suite Connect to Your LinkedIn Ad Accounts?

Every LinkedIn skill in the BlueAlpha Marketing Plugin is backed by an MCP connector that gives Claude direct, permissioned access to your LinkedIn Ads account data. MCP-native means the whole system lives inside Claude, Codex, or any AI workspace your team already uses. No separate dashboard to check, no CSV export to email around.

The BlueAlpha LinkedIn connector gives Claude the ability to:

• List every LinkedIn ad account the connected token can access (across business managers)

• Pull campaign-level configuration: status, daily budget, pacing strategy, targeting criteria, cost type, optimization target, creative selection mode

• Read creative-level data: review status, serving status, intended status, Lead Gen Form references, content references

• Run analytics by single pivot (CAMPAIGN, CREATIVE, MEMBER_SENIORITY, MEMBER_JOB_FUNCTION, MEMBER_INDUSTRY, MEMBER_COMPANY_SIZE, and more) over any date range

• Run multi-pivot statistics finders for more nuanced segment slicing

• Detect serving statuses and holds (AUDIENCE_COUNT_HOLD, BILLING_HOLD, CAMPAIGN_GROUP_STATUS_HOLD) directly from the API

The connector handles authentication, account access, and field-level permissions. One sign-in, no API keys to share, no ad account IDs to copy-paste.

How to Install the BlueAlpha LinkedIn Plugin

There are two pieces to install: the BlueAlpha MCP connector (so Claude can read your LinkedIn ad accounts) and the BlueAlpha Marketing Plugin itself (the skills that put that data to work). Total setup time is about a minute.

Step 1: Install the BlueAlpha MCP connector

1. Open Settings in the Claude desktop app

2. Go to Connectors > Add custom connector

3. Name it: BlueAlpha MCP

4. URL: https://mcp.bluealpha.ai/mcp

5. Click Connect and sign in with your BlueAlpha account

That single sign-in wires Claude up to your LinkedIn ad accounts (and your Google Ads, Meta, and Meridian models if you have them). No keys, no IDs, no config files.

Step 2: Install the plugin

Option A: Cowork (drag-and-drop)

1. Go to github.com/bluealpha-labs/bluealpha-plugins

2. Click Releases on the right rail and open the latest release

3. Expand Assets and click bluealpha-marketing-plugin.plugin to download

4. Drag the downloaded file into an open Cowork session and click Install when prompted

Option B: Claude Code (slash commands)

/plugin marketplace add <https://github.com/bluealpha-labs/bluealpha-plugins.git>

/plugin marketplace add <https://github.com/bluealpha-labs/bluealpha-plugins.git>

Step 3: Try a skill

"Run auto-optimize on my LinkedIn account 515970088 and tell me what's blocking delivery."

LinkedIn Skill Reference: Prompts and Real Output

Every skill is invoked with a sentence. Below is a copy-paste prompt for each one, plus the real output from a run on the BlueAlpha account (account 515970088, USD, 4 active campaigns, $22,546 of spend over the trailing 90 days).

linkedin-auto-optimize

Run auto-optimize on my LinkedIn account 515970088. Tell me whether any active campaigns are blocked from delivering, whether pacing is healthy, and what the account health score is.

The skill produces an overall 0-100 health score plus four sections: blockers (resolve first), pacing issues, structure hygiene, and routing recommendations to the deeper-dive skills.

Real example from BlueAlpha's account: the model scored 65/100. Active delivery was healthy ($300/day budget at 100% utilization), but Traffic-Retargeting was under-pacing at 6% utilization ($3/day actual vs $50/day budgeted) and 6 paused campaigns had outstanding count-hold blockers totaling $410/day of paused delivery that would block resumption.

linkedin-performance-digest

Generate the weekly performance digest for my LinkedIn account 515970088. Trailing 14 days vs prior 14 days. Format: headline, what happened, what drove it, what to watch.

A single-page narrative report: spend, clicks, LP clicks, leads with period-over-period deltas, named-driver attribution, and a what-to-watch section.

Real example: trailing 14 days vs prior 14 days showed spend +55% ($2,528 vs $1,626) and landing-page clicks +180% (84 vs 30). CPC moved down to $4 from $5 and cost per landing-page click compressed from $54 to $30. The skill named Traffic - Image - HubSpot targeting as the driver: new in the window, $826 spent, 33 LP clicks at 100% click-to-LP rate.

linkedin-demographic-deep-dive

Run the demographic deep-dive on my LinkedIn account 515970088, trailing 90 days. Tell me where my spend is leaking outside the intended targeting.

Four-section report: who you're paying to reach (intended vs actual), who's actually engaging, where you're wasting, and where to expand, across seniority, job function, company size, and industry pivots.

Real example: BlueAlpha's account had zero conversions tracked across $22,546 of spend. The LinkedIn Insight Tag isn't recording conversions, so the optimizer is bidding on clicks, not outcomes. Action #1: install the Insight Tag. Next biggest finding: a cross-pivot SMB cluster (Owner/Partner seniorities at 17% LP-click rate AND 2-10/11-50 employee companies at 18% LP-click rate), engaging at ~2x the account average (9.4%). Marketing (the intended target) got 47% of spend but 53% leaked to unintended functions, chiefly Business Development ($3,104 at 13.4% LP-click rate, a sleeper to add) and Engineering/Product Management ($965 at ~5% LP-click rate, waste to cut).

linkedin-audience-health-check

Audit the matched-audience health on my LinkedIn account 515970088. Find delivery-blocked campaigns, single-point-of-failure audiences, and any audiences shared across active campaigns.

Five sections: delivery-blocked campaigns (with audience URNs and daily budget at stake), audience usage map (risk-ranked), stale/declining audiences, targeting layer breakdown, and a manual validation checklist.

Real example: 3 campaigns delivery-blocked by AUDIENCE_COUNT_HOLD totaling ~$410/day of unblockable delivery. The audience usage map flagged urn:li:adSegment:64784133 as a compound-risk audience: shared between Engagement-image and Engagement-video (auction overlap), AND the sole matched audience for Engagement-video (single point of failure).

linkedin-targeting-overlap-finder

Find the campaigns on my LinkedIn account 515970088 that are bidding against each other because their targeting overlaps. Compute Jaccard similarity per facet.

Ranked list of overlapping campaign pairs above the 60% threshold, with per-facet Jaccard breakdown, daily-budget estimate of auction-overlap dollars, and a specific recommendation per pair.

Real example: 2 overlapping pairs detected. Engagement-image / Traffic-HubSpot targeting at 70% overlap: estimated $35/day = ~$1,050/month in auction competition. Engagement-image / Engagement-video at 62% overlap: estimated $31/day = ~$930/month. Total: ~$2K/month.

linkedin-creative-fatigue-watchdog

Score the creative fatigue on my LinkedIn account 515970088. Use objective-appropriate metrics — CTR for Engagement campaigns, LP-click rate for Traffic. Surface fatigued creatives, weak-from-launch creatives, and top performers.

Picks the right primary metric per campaign objective (CTR for Engagement, LP-click rate for Traffic), then classifies each creative as fatigued, weak-from-launch, waste, or top performer.

Real example: 17 creatives in active campaigns, 4 with sufficient volume to score. Zero false fatigue alerts. Most of the active creative rotation was too new for fatigue scoring. Creative naming hygiene finding: 8 of 17 active creatives have empty name fields.

linkedin-bid-strategy-audit

Audit the bid configurations on my LinkedIn account 515970088. Flag cost-type mismatches, suspicious unit costs, and missing optimization targets.

Checks each campaign's costType, unitCost, optimizationTargetType, creativeSelection, and pacingStrategy against the campaign's objectiveType.

Real example: No cost-type mismatches and no suspicious bid amounts. But all 4 active campaigns are using ROUND_ROBIN creative selection instead of OPTIMIZED. Switching typically yields 15-30% engagement improvement on multi-creative campaigns.

linkedin-frequency-saturation-report

Audit the frequency caps on my LinkedIn account 515970088. Flag campaigns running without caps on matched audiences — the highest saturation risk.

Audits frequencyOptimizationPreference per campaign, classifies the audience type, and flags mismatches.

Real example: 3 of 4 active matched-audience campaigns have no frequency cap set. The skill correctly skipped the retargeting campaign (dynamic segments, higher frequency is intentional) and recommended caps of 3-5 per 7 days for the matched-audience campaigns.

linkedin-lead-form-quality-auditor

Audit my LinkedIn Lead Gen Forms on account 515970088. Find creatives spending without producing leads, declining volume, and orphan forms.

Identifies every creative with leadgenCallToAction set, classifies each as spending-without-leads / declining / dormant / healthy, computes a cost-per-lead leaderboard, and maps every form URN to the creatives using it.

Real example: 2 Lead Gen creatives in active campaigns. No spend-without-leads cases, no declining volume. But one orphan form detected: only referenced by paused creatives, an archive candidate.

linkedin-full-monty

Run the full monty on my LinkedIn account 515970088. Compose the structure audit, audience health, overlap, frequency, creative fatigue, lead forms, demographic deep-dive, bid audit, and performance digest into one report.

Runs all nine sub-skills against shared data (pulled once at the start) and composes the findings into a single 4-8 page audit with cross-skill prioritization.

Real example, cross-prioritized Top 5 actions for BlueAlpha:

1. Install the LinkedIn Insight Tag: zero conversions tracked across $22K of spend

2. Decide on the SMB cluster: $3,808/quarter to small-company / founder audiences engaging at 2x the account average

3. Resolve count holds on the paused Lead Gen and Conversions campaigns: $410/day of paused delivery

4. Address the Engagement-image / HubSpot-targeting overlap: $35/day of auction competition

5. Switch all 4 active campaigns from ROUND_ROBIN to OPTIMIZED creative selection + set frequency caps

Why BlueAlpha LinkedIn Skills Beat Another LinkedIn Dashboard

There are two kinds of products in LinkedIn ad management, and BlueAlpha is neither.

The first is the platform itself: Campaign Manager. It surfaces metrics and lets you configure campaigns, but it doesn't make decisions for you. Every report you pull is a starting point for another conversation, not a recommendation. LinkedIn's UI is famously dense: even experienced operators miss count holds, ROUND_ROBIN settings, and audience overlap because the dashboards don't surface them.

The second is the standard agency reporting stack: a weekly slide deck with screenshots from Campaign Manager pasted in, the same five vanity metrics every week, no specific actions.

BlueAlpha is the third option: causal measurement underneath, real agents producing deployment-ready specs, MCP-native. Not a measurement vendor. Not another dashboard. The decision layer underneath your LinkedIn account.

The difference shows up in results. When BlueAlpha ran incrementality tests for Cann, AppLovin's own dashboard swung between 2x and 16x ROAS and the third-party tool Triple Whale showed 4.77x, but the measured incremental lift was zero, saving Cann $480K per year in misallocated spend. For Klover, BlueAlpha's measurement cut Meta iOS spend by 50% with zero lost conversions, freeing budget to scale Apple Search Ads by 10x. For beehiiv, causal measurement caught a channel overreporting performance by 345%. For Pettable, nine months of causal measurement reduced blended CAC by 14%, unlocking $2.12M in annualized savings.

FAQ

Do I need to give BlueAlpha admin access to my LinkedIn account?

No. The connector uses LinkedIn's standard OAuth flow with read scopes. BlueAlpha can read your campaign configuration, creative data, and analytics, but cannot create, modify, or pause campaigns. All recommendations are surfaced to you for execution.

What if I run LinkedIn ads across multiple accounts?

The connector lists every ad account the connected token can access. The skills accept an account_id parameter, so you can scope each invocation to one account.

Will the skills make changes to my LinkedIn account automatically?

Not currently. The LinkedIn skills produce recommendations, refresh queues, exclusion lists, and configuration changes for you to review and apply. On-platform execution is on the roadmap and will be opt-in per-account.

Some skills mention API constraints. What's the impact?

LinkedIn's API exposes most but not all of the data we'd want. Three skills work as configuration audits + indirect signals because of API gaps: linkedin-audience-health-check can't see audience sizes (only campaign-level holds), linkedin-frequency-saturation-report can't see unique reach (so it's a cap-config audit, not a true frequency measurement), and linkedin-lead-form-quality-auditor can't see form-level submission data (only creative-level lead counts). Each skill surfaces the limitation explicitly and produces a manual-validation checklist. All 10 skills produce real, actionable findings today.

How often should I run each skill?

linkedin-auto-optimize and linkedin-performance-digest: weekly.
linkedin-creative-fatigue-watchdog and linkedin-audience-health-check: bi-weekly.
linkedin-targeting-overlap-finder, linkedin-frequency-saturation-report, and linkedin-bid-strategy-audit: monthly.
linkedin-demographic-deep-dive, linkedin-lead-form-quality-auditor, and linkedin-full-monty: quarterly.

Can the skills work without conversion tracking?

Yes, but with a caveat. When the LinkedIn Insight Tag isn't installed, the skills fall back to landing-page-click rate as the performance proxy. They surface this limitation explicitly and rank "Install the Insight Tag" as the #1 action.

Do the LinkedIn skills work with the Google Ads, MMM, and Meta skills?

Yes. They ship in the same plugin (bluealpha-marketing-plugin) and are designed to compose.
The natural cross-platform planning loop is mmm-health-check > mmm-trust-router > linkedin-auto-optimize (and equivalents for Google / Meta) > mmm-scenario-planner > channel-specific execution.

Where can I find the full setup documentation?

The complete installation walkthrough, authentication scopes, troubleshooting, and the full tool reference live on the BlueAlpha MCP documentation page.

How do I prove LinkedIn ads actually drive B2B pipeline?

The demographic deep-dive skill cross-references who you're paying to reach against who's actually engaging, broken down by seniority, job function, company size, and industry. When the LinkedIn Insight Tag is installed, conversion data flows into the analysis. When it's not (as was the case in BlueAlpha's own account audit), the skill falls back to landing-page-click rate as a proxy and ranks "Install the Insight Tag" as the #1 action. The full-monty skill then composes all ten audits into a single report with dollar-denominated findings your CFO can read.

Your next step

Pick the skill that matches the question you're trying to answer today, copy the prompt above, and run it in Claude. If you're new to the LinkedIn skills, start with linkedin-auto-optimize. It's the right first move on any account you haven't audited, and it produces the health score that every downstream skill's caveats hang on.

If the account scores below 85, linkedin-audience-health-check is the second prompt to run. It identifies which audiences are causing the structural issues, the #1 cause of "why isn't my LinkedIn delivering?" on a B2B account.

If the account looks structurally fine, the highest-leverage planning skill is linkedin-demographic-deep-dive. It surfaces the targeting drift dollars that compound over time and gives you the specific job functions, seniorities, and company sizes to add or exclude.

If you haven't installed the plugin yet, setup takes about a minute. To see the full platform, visit the product page. Browse the full prompt library for prompts across all channels. If you'd rather see the skills run on your account before installing, book a walkthrough.

The same plugin ships Google Ads skills, Meta Ads skills, TikTok Ads skills, and Meridian MMM skills. For a side-by-side comparison of all MCP servers, see the full MCP comparison.