Module 04 · Profiles and person signals

Outcome: an opening line based on something the person actually said, plus a check on whether they just moved seats.

Surface: MCP server
Level: Beginner
Uses: linkedin_profile_enrichlinkedin_profile_postslinkedin_profile_reactionslinkedin_profile_commentscheck_job_changecheck_promotions
Credits: ~5.7 to run every example once
Prerequisite: A LinkedIn profile URL

Tools in this module

Tool	Cost	Required input	Returns
`linkedin_profile_reactions`	0.2	`profile_url`	Posts they reacted to
`linkedin_profile_comments`	0.2	`profile_url`	Comments they left, with parent-post context
`linkedin_profile_posts`	0.3	`profile_url`	Posts they authored
`linkedin_profile_enrich`	1	`profile_url`	Work history, education, skills
`check_job_change`	2	`profile_url`	Whether they changed jobs recently
`check_promotions`	2	`profile_url`	Whether they were promoted recently

The three activity tools cost 0.2–0.3 credits. They are the cheapest personalization on the server and the most under-used.

Example prompts

linkedin_profile_enrich — the record


Use linkedin_profile_enrich with profile_url
"https://linkedin.com/in/janedoe".

Return current title, current company, time in role, previous two roles,
and top five skills.

linkedin_profile_posts — what they say


Use linkedin_profile_posts with profile_url
"https://linkedin.com/in/janedoe", max_posts 20.

Return each post's date, first 200 characters, and engagement count.
Rank by engagement.

linkedin_profile_reactions — what they endorse

Two filters worth knowing: posted_limit accepts any, 24h, week, month, 3months, 6months; max_results is 1–50 and defaults to 3.


Use linkedin_profile_reactions with profile_url
"https://linkedin.com/in/janedoe", posted_limit "3months",
max_results 25.

Return the post author, the topic and the post URL.
Tell me the three topics that come up most.

linkedin_profile_comments — their own words


Use linkedin_profile_comments with profile_url
"https://linkedin.com/in/janedoe", posted_limit "6months",
max_results 25.

Return the comment text and what the parent post was about.
Flag any comment that names a tool or a vendor.

Comments beat posts for personalization. A post is marketing; a comment is an opinion. At 0.2 credits it is the cheapest useful signal on the server.

check_job_change — the timing signal


Use check_job_change with profile_url
"https://linkedin.com/in/janedoe".
If they changed jobs, tell me the new company and roughly when.

check_promotions — the budget signal


Use check_promotions with profile_url
"https://linkedin.com/in/janedoe".
If promoted, tell me the new title and roughly when.

The 1.7-credit personalization pass

Pull the cheap signals first

Reactions and comments at 0.2 each, posts at 0.3. Total 0.7 for a picture of what they care about.

Read before you enrich

If nothing in the activity is relevant, stop. You just saved a credit.

Enrich the ones worth writing to

linkedin_profile_enrich at 1 credit fills in the history and skills.

Check timing only on the shortlist

check_job_change and check_promotions are 2 credits each. Run them on people you are about to contact, not on a list.

Run it as one message:


For https://linkedin.com/in/janedoe:

1. linkedin_profile_comments, posted_limit "3months", max_results 20
2. linkedin_profile_posts, max_posts 10
3. linkedin_profile_enrich

Then draft a two-sentence opener that quotes one specific thing they said.
Do not write anything generic — if there is no usable material, say so.

Workflow prompts

Person signals are worth most when they reach you without being asked for.

Alert the owner when a champion moves

take the champions list out of the CRM
check each for a job change
alert the rep who owns them and open a task


1. Pull every Attio contact tagged "champion" that has a LinkedIn URL.
2. For each: check_job_change — 2 credits per person, so tell me the
   headcount and total cost before starting.
3. For anyone who moved: get the new company and role.
4. Post one Slack message per mover to #sales:
   "<name> moved from <old company> to <new company> as <new title>.
   Owner: <record owner>."
5. Create an Attio task for that owner, due in 3 days, titled
   "Reach out — <name> just moved to <new company>".

Personalize a sequence from what people actually said

take the list about to be sequenced
pull their comments and posts
write one custom line per person, or mark them skip


For each of these 20 LinkedIn URLs:

1. linkedin_profile_comments — posted_limit "3months", max_results 20
2. linkedin_profile_posts — max_posts 10

Then write one opening line per person that references something specific
they said, and include the URL of the post or comment it came from.

If a person has no usable material, write "SKIP — no signal" instead of
inventing something. I would rather send 12 good lines than 20 generic ones.

Return a CSV: name, linkedin_url, opening_line, source_url.

Brief before a call, from the calendar

read today’s external meetings
pull the attendee’s recent activity
hand back a short brief per meeting


1. List my external calendar meetings for today with attendee names
   and email domains.
2. For each external attendee, find their LinkedIn profile, then run
   linkedin_profile_enrich and linkedin_profile_comments
   (posted_limit "3months").
3. Give me one brief per meeting: who they are, how long in seat, what
   they have been talking about, and one question worth asking them.
Keep each brief under 120 words.

Best prompting practices

Use posted_limit. Without it the default is any and you get stale material. 3months is usually right.
Raise max_results. The default is 3 on reactions and comments. At 0.2 credits per call, 25 costs the same as 3.
Ask for the source URL. A quote you cannot link back to is a quote you cannot check.
Say “if there is no usable material, say so.” Otherwise you get a generic opener dressed as personalization.
Run the 2-credit signal tools last. Job change and promotion matter for timing, not for qualification.

Where this breaks

A profile with no public activity returns empty and the model may fill the gap from the profile headline. Ask for the post URL alongside every quote — no URL, no quote.

check_job_change on a whole list is 2 credits a row. For “who moved recently” across many companies, use find_people with recently_changed_jobs: true instead — see module 02.

Further automation

MCP Profile Signals is this module as a skill file — /mcp-profile-signals pulls the cheap signals, decides whether to enrich, and drafts the opener with its source URL.

Job Change Detector and the champion-tracking agents run it on a schedule, so a move reaches you before the sequence does.

Browse all GTM agents →

Next module

05 — Engagement and social sourcing, the other half of the social-signals cluster: turning a post into a warm list.

Reference: LinkedIn Profile Enrich · Profile Posts · Profile Reactions · Profile Comments · Check Job Change · Check Promotions · Job Changes & Champions