Megan wants to start medical volunteering this year as a part of her 2026 goals. Instead of endless Googling, I built her an interactive ranking page in about 27 minutes of AI churning with only 3 interactions on my part, its pretty good:
see it here:
https://www.cosgrove.blog/mission-trips/
What follows was written by the same task:
What It Does
A single HTML file with 100+ ranked opportunities, filterable by country,
duration, and confirmed dates. Each card shows PA-specific policies,
costs, charity ratings, and impact metrics. Top pick: Refuge
International in Guatemala with a $750 scholarship for licensed
providers.
How It Was Built
Used Claude Code (Anthropic’s CLI) with Opus 4.5. Total usage:
~36k tokens. Claude kicked off hundreds of searches, generated HTML, explained deployment options,
and provided upload to this wordpress instance on AWS.
Embedded on WordPress via iframe. Done.
Bonus: Ralph Wiggum Plugin
Claude Code supports plugins like https://github.com/anthropics/claude-co
de/blob/main/plugins/ralph-wiggum/README.md — an autonomous iteration
loop. Run /ralph-loop “build X” –completion-promise “DONE” and Claude
keeps working until success, seeing its previous attempts in git history.
task:
/ralph-wiggum:ralph-loop ‘
## TASK: Build an HTML Research Directory of PA-Friendly Medical Mission Opportunities in Latin America
Create a comprehensive, browsable HTML website documenting medical mission volunteer opportunities that meet ALL of these criteria:### REQUIRED CRITERIA (ALL must be met for inclusion):1. **Role**: Must explicitly accept Physician Assistants (PAs) – not just MDs/DOs or RNs2. **Location**: Spanish-speaking Latin America ONLY (Mexico, Guatemala, Honduras, El Salvador, Nicaragua, Costa Rica, Panama, Colombia, Venezuela, Ecuador, Peru, Bolivia, Chile, Argentina, Uruguay, Paraguay)3. **Duration**: Programs offering 2-6 week volunteer commitments4. **Scope**: General/family medicine focus (not surgical specialty only)5. **Not Student Rotations**: this search is focused on programs for career professionals, not those currently working to become PAs but those that are already PAs### SEARCH THESE ORGANIZATION TYPES:– University medical schools and PA programs with global health initiatives– Faith-based/church medical mission organizations– Hospital systems and large medical groups with volunteer programs– NGOs and nonprofits (like Remote Area Medical, Global Brigades, etc.)– International health organizations accepting mid-level providers### FOR EACH OPPORTUNITY FOUND, CAPTURE:– Organization name and website URL– Specific program name (if applicable)– Countries/regions served– Trip duration options– Whether PAs are explicitly mentioned as eligible– Application process/requirements– Approximate costs (if available)– Contact information– Notes on experience level required### OUTPUT REQUIREMENTS:Create these files in the project directory:1. **index.html** – Main browsable site with:– Clean, professional design– Filterable/searchable table of all opportunities– Filters by: country, duration, organization type, cost range– Cards or list view of each opportunity– Links to detailed pages2. **opportunities.json** – Structured data file with all findings3. **sources.md** – List of all URLs searched and date accessed### ITERATION STRATEGY:– Iteration 1-5: Search and document university-based programs– Iteration 6-10: Search faith-based/church organizations– Iteration 11-15: Search hospital systems and medical groups– Iteration 16-20: Search NGOs and nonprofits– Iteration 21-25: Build HTML site, deduplicate, verify links, finalize### COMPLETION CRITERIA:When you have:– Researched at least 200 potential organizations– Documented at least 100 verified PA-eligible opportunities– Built a working HTML site with all data– Created the JSON data file– Listed all sourcesOutput: <promise>RESEARCH_COMPLETE</promise>### IF STUCK AFTER 20 ITERATIONS:– Document what searches were attempted– List any organizations that need manual verification– Complete the HTML with whatever valid data was found– Note gaps in research for manual follow-up– Output the completion promise anyway–max-iterations 25 –completion-promise “RESEARCH_COMPLETE’