Motivation - Holes are good. They let you go places.

Method 1

Download Blender 2.93. Do NOT use the Windows Store Version!
Download Metaforce and extract your ISO. This can take 1hr+. Do not use an extraction filepath longer than what is shown below:

Find the .blend file for the area you would like to analyze and open it in blender. e.g. MP1\Metroid1\!1IntroLevel1027_158EFE17\00 Exterior Docking Hangar\!area_D1241219.blend
Uncheck the Render group:

This is stuff you can see in-game but not neccessarily collide with.

Ensure the Collision/CMESH is checked and the hidden "eye" is open.

This is just the collision data.

In the modelling workspace, select all "Non-Manifold" edges:

Switch to "Edge Select":

Single-click one of the non-manifold edges and press g to distort it around and look for holes:

Once you find a hole worth investigating, you can view it's coordinates by changing the selection mode to Vertex, selecting a vertex adjacent to the hole, then opening the transform window by pressing n. You will need to switch the coordinate view to Global for the coordinates to make any sense.

Rinse and repeat until you find a real stinky one.

Method 2

Don't use this method, it is inferior and overly-acedemic. This information is being preserved for posterity only.

Repeat steps 1-5 from Method 1
Save the file as pre.blend
Select all vertices and apply "Merge by Distance:"

Re-apply the "Merge By Distance" with 0.1m:

Save the file as post.blend
Using these commands:

 blender --background post.blend --python test.py > post.txt
 blender --background pre.blend --python test.py > pre.txt

run this python script:

import bpy

found_it = False
for ob in bpy.context.scene.objects:
    if ob.data.name != "CMESH":
        continue

    print("location:" + str(list(ob.location)))
    mesh = ob.data
    for vertex in mesh.vertices:
        print(list(vertex.co))

    assert(not found_it)
    found_it = True

assert(found_it)

Run this python script:

import math

def distance(a, b):
    return math.dist(a,b)

def to_vec3(string):
    string = string.replace("[", "").replace("]", "").replace("location:", "").replace(",", "")
    data = [float(x) for x in string.split(" ")]
    return data

def parse_file(filename):
    with open(filename, "r") as file:
        text = file.read()
        lines = text.split("\n")
        location = to_vec3(lines[5])
        return(lines[5:], location)

(pre, pre_location) = parse_file("pre.txt")
(post, post_location) = parse_file("post.txt")

diff = list()
for line in pre:
    if line not in post:
        diff.append(line)

for line in post:
    if line not in pre and line not in diff:
        diff.append(line)

# print(diff)
assert pre_location == post_location
location = pre_location

diff_coords = list()
for line in diff:
    coord = to_vec3(line)
    coord = [location[0] + coord[0], location[1] + coord[1], location[2] + coord[2]]
    diff_coords.append(coord)

for coord_a in diff_coords:
    for coord_b in diff_coords:
        if coord_a == coord_b:
            continue

        dist = distance(coord_a, coord_b)
        if dist < 5.0:
            diff_coords.remove(coord_b)

for coord in sorted(diff_coords):
    print("%0.3f, %0.3f, %0.3f" % (coord[0], coord[1], coord[2]))

Inspect the printed coordinates for holes in your favorite hole-viewing program

toasterparty/finding-holes-mp1.md

Method 1

Method 2