gc-alumni

name	gc-alumni
description	Use for questions about Clemson Graphic Communications alumni outcomes — where graduates work now, first-job placement, career trajectories, AI exposure of GC occupations, industry/occupation distribution, grad-school rates, and outcome trends by graduation cohort. Data comes from the gc-alumni MCP server (read-only SQLite). For degree rules or course content use gc-advisor or clemson-curriculum instead.
metadata	{"version":"1.0.0"}

GC Alumni Outcomes

Query the Clemson GC alumni database via the gc-alumni MCP server (read-only). Never answer from memory — pull from the DB.

Tools

Tool	Use for
search_alumni	Look up a specific graduate by name → grad year, first job, current job, AI trajectory
get_alumni_photo	Return a graduate's LinkedIn profile photo — see Photos below
outcomes_by_year	Post-grad outcome mix (employed / grad school / seeking) and first-job AI exposure by cohort
occupation_distribution	First-job SOC occupations ranked by count, with AI exposure score and BLS trajectory
industry_distribution	First-job NAICS industries ranked by count
review_inbox	Quick count: open student requests + flagged records awaiting admin attention
pending_requests	Full list of open student bug/feature/data requests with AI triage (category, summary, severity, effort)
query	Any ad-hoc read-only SQL (SELECT / WITH only); results capped at 200 rows by default
get_schema	Inspect a table's CREATE SQL before writing a query (table="" returns all)
list_tables	See all available tables and views

Key tables and views

Start here for most questions:

• v_alumni — denormalized view: person + current job + first job with AI exposure and trajectory. Best for browsing or filtering alumni.
• v_ai_quadrant — trajectory label per SOC occupation: Opportunity, Adapt, Extinction risk, Resilient, Stable, or Declining.

Core tables:

• person — one row per graduate: grad_year, grad_term, post_grad_outcome (employed | grad_school | seeking | unknown), current_company, current_title, current_location, went_to_grad_school
• position — full job history: is_first_job, is_current, soc_code, naics_code, ai_impact (1–10), future_difficulty (1–10)
• ai_exposure — per SOC code: exposure_score (0–10, higher = more AI-exposed), outlook (BLS OOH growth text), median_pay
• education — is_graduate=1 flags post-bachelor degrees (Master's, PhD, MBA)
• market_segment — sector label on each person (Print & Packaging, Agency, CPG, …)

Useful joins:

-- Alumni with first job and AI exposure
SELECT full_name, grad_year, first_job_title, first_job_company,
       first_job_ai_exposure, first_job_ai_trajectory, current_title, current_company
FROM v_alumni
WHERE grad_year BETWEEN 2020 AND 2024
ORDER BY grad_year, full_name;

-- Occupations by AI risk
SELECT occupation, exposure_score, outlook, trajectory
FROM v_ai_quadrant
ORDER BY exposure_score DESC;

Photos

get_alumni_photo returns two things: an inline image (you see it visually) and a text block with a photo_url and filename. Use those to deliver the photo:

1. Call get_alumni_photo(who) — note the photo_url and filename from the text response.
2. Call fetch_url_to_workspace(url, filename) — downloads and saves the photo to /workspace/agent/.
3. Call send_file(path=filename, text="Here's [name]'s photo") — delivers it to the user.

Do NOT say "I've sent you the photo" until after send_file succeeds.

Rules

• search_alumni first for name lookups — it's faster and avoids SQL for simple queries.
• outcomes_by_year / occupation_distribution / industry_distribution first for aggregate/trend questions before writing a custom query.
• get_schema before complex queries — confirm column names, especially for position (many flags) and v_alumni (rich join).
• The query tool is SELECT/WITH only; writes are blocked at the DB level.
• post_grad_outcome on person is normalized (employed | grad_school | seeking | unknown). The raw status column holds the original text from the grad-seniors survey.
• is_first_job=1 on position marks the first post-graduation job; is_current=1 marks the current role. A person can have multiple positions — filter by flag, don't assume one row per person.
• AI exposure scores come from the Karpathy/ONET OOH→SOC crosswalk; v_ai_quadrant combines exposure + BLS growth outlook into the trajectory label.
• Default limit for query is 200 rows. Pass a higher value if aggregates might be truncated.