Extremely lean and efficient. No explanation of what pytest is or how Python works. Every section delivers actionable information Claude wouldn't already know. The cross-reference to java-test-updater avoids repetition.

Provides executable code examples throughout (conftest.py shim, pytest commands, concrete test rewrites). The failure taxonomy table maps specific error messages to specific actions. Copy-paste ready examples for mock decoupling and assertion fixes.

Clear multi-step workflows with explicit validation: snapshot update workflow has numbered steps with review-before-update checkpoint. The 'Do not' section provides explicit guardrails. Output format creates a structured triage checklist with classification steps.

Content is well-organized with clear sections and tables, but everything is inline in one file. The reference to java-test-updater is good, but the skill could benefit from splitting detailed examples (like the conftest shim patterns) into a separate reference file for complex migrations.

Lists multiple specific concrete actions: 'triaging assertion failures from mock-coupling failures from genuine regressions' and 'using Python's introspection to automate where safe'. Describes concrete decision-making (update vs. fix vs. delete).

Clearly answers both what ('Updates broken pytest tests...triaging...using Python's introspection') AND when ('Use when a refactor or API change leaves a pile of failing tests and you need to decide update vs. fix vs. delete').

Includes natural keywords users would say: 'pytest', 'tests', 'refactor', 'API change', 'failing tests', 'assertion failures', 'mock'. These are terms developers naturally use when dealing with broken tests.

Highly specific niche: pytest test repair after intentional code changes. The combination of 'pytest', 'refactor/API change', and 'triaging failure types' creates a distinct trigger profile unlikely to conflict with general testing or coding skills.