Grasp · Comprehension Index Scan Report

Who still understands anthropics/anthropic-sdk-python?

Computed from git metadata only; no source code contents were read.
touch-weighted · 3-year history window · last commit 17d ago · scores are comparable only against same-mode scans

Grasp Score

touch-weighted living knowledge across the repo

Active humans

13 / 61

contributors active in last 90 days vs. all-time

AI-attributed minimum

≥ 87.2%

hard evidence only. Just 82% of commits here carry any attribution signal, so the real share is higher

Inactive share

7.9%

work by authors with no commit in the window; silence, not proof of departure

Composition

72/13/10/5%

source / tests / docs / config share of scored work

Living, source only

14%

the headline recomputed over hand-written source alone

Score stability

stable

90-day-cliff variant reads 10%–16% across 60/90/180-day activity windows; read the width before the middle value

Dark-code ranking

Modules ranked by comprehension risk: the unknown share, amplified by ongoing change and thin ownership. The green meter is living knowledge.

Module	Risk	Living knowledge	AI-attr.	Inactive	Bus	Heat/180d
src/anthropic	CRIT 95	5%	94%	3%	1	120
(root)	CRIT 93	7%	92%	6%	1	369
src/anthropic/types	CRIT 88	12%	91%	7%	1	1222
src/anthropic/resources	CRIT 84	16%	82%	4%	1	187
scripts	HIGH 79	7%	92%	3%	2	37
src/anthropic/_utils	HIGH 78	4%	97%	6%	1	12
.github	HIGH 76	9%	88%	13%	1	18
examples	HIGH 75	25%	70%	23%	1	57
tests	HIGH 74	20%	77%	12%	2	376
src/anthropic/lib	MED 52	43%	48%	18%	2	196

Public-repo caveat: "departed" includes drive-by contributors, overstating attrition relative to a company repo with a real roster. Engagement is a proxy for comprehension. Treat it as a potential indicator rather than a diagnosis.