The body is a lean, dense audit checklist with no padding of concepts Claude already knows; the longer routing-rationale paragraph covers non-obvious architecture rather than fluff, so it sits at the level-3 'every token earns its place' anchor (borderline with 2 but justified).

Provides an executable register_routes snippet, exact commands ('python manage.py find_enum_collisions', 'hogli build:openapi'), and concrete decorator names — fully actionable, copy-paste-ready guidance rather than the abstract level-1 anchor.

Sequences triage-first (check generated output) then a numbered 1–16 checklist with a decision flowchart and explicit validation steps ('run hogli build:openapi and verify generated MCP tool schemas'), matching the level-3 anchor with checkpoints and a feedback loop.

A concise overview points to four well-signaled, one-level-deep reference files (all verified present, including the #choicefield--explicit-choices anchor), with detail appropriately split out — exactly the level-3 anchor.

Lists multiple specific concrete actions — 'editing, reviewing, or auditing DRF viewsets and serializers' and 'field typing, schema annotations, enum collision fixes, and OpenAPI spec quality' — matching the level-3 anchor rather than the partial level-2 anchor.

Explicitly answers both what ('Covers field typing, schema annotations, enum collision fixes, and OpenAPI spec quality') and when ('Use when editing, reviewing, or auditing...'), matching the level-3 anchor with an explicit trigger clause.

Natural domain terms a PostHog developer would say ('DRF viewsets', 'serializers', 'OpenAPI spec', 'rest_framework') plus explicit path/file triggers give good coverage, beyond the level-2 'some relevant keywords' anchor.

Scoped to PostHog-specific paths (posthog/api/, products/*/backend/api/) and gated on 'rest_framework' imports, giving it a clear niche unlikely to fire for unrelated skills.