Conciseness

The body is lean and policy-dense with no explanation of concepts Claude already knows; the opening address-book analogy ("A person doesn't bulk-load their whole address book into working memory…") is rhetorical framing that could be trimmed to one line. Not level 5 because those few motivational lines don't strictly earn their place; not level 3 because the padding is minor and everything else is tight.

Actionability

Concrete, executable guidance throughout: `get_page <slug>` (MCP), pulling `graph` / `backlinks`, and a 3-level escalation ladder with explicit stop conditions ("stop there", "only when relationship context is needed"). Not level 5 because the exact CLI command forms (`gbrain graph`, `gbrain backlinks`) live only in a frontmatter comment rather than the body, leaving a minor gap for non-MCP use.

Workflow Clarity

The WHEN-to-retrieve conditions, skip rules, and the numbered escalate-and-stop ladder form a clear, unambiguous sequence with built-in checkpoints ("stop there", "no bulk-loading"). Not level 5 because the framing "Encode that reflex" sits in unresolved tension with "Judgment first — retrieve when it changes the quality of the reply, not reflexively", leaving slight ambiguity about whether to act reflexively or deliberatively; not level 3 because the sequence and stop conditions are explicit, and the skill is read-only so no validation loop is required.

Progressive Disclosure

No bundle files exist and the skill is under 50 lines with well-organized sections (Trigger policy, Retrieval spec, failure rationale, See also). Per the rubric's simple-skill guidance, well-organized sections with no need for external references score 5; the "See also" pointers to sibling skills are one level deep and clearly signaled.

Specificity

Names the domain (memory retrieval) and one concrete action ("open the brain page for a salient entity before answering from memory"), but stops there. It is not level 4 because only a single action is given and "When/what to retrieve" gestures at scope rather than listing actions.

Completeness

The "what" is clear (open the entity's brain page before answering), but the "when" is only weakly implied by "When/what to retrieve" and "before answering from memory"; there is no explicit "Use when…" clause, which caps completeness at 3. Not level 2 because a discernible trigger condition is present alongside the clear what.

Trigger Term Quality

Relevant keywords are present ("retrieve", "brain page", "entity", "answering from memory") but they are internal jargon; natural user phrases like "who is" or "tell me about" appear only in a frontmatter comment and are explicitly routed to a different skill. Not level 4 because common natural-language variations users would actually say are missing from the description itself.

Distinctiveness Conflict Risk

"open the brain page for a salient entity before answering from memory" carves out a niche, but it overlaps heavily with the sibling skills/query search skill and the description does not disambiguate when this reflex applies versus an explicit search request. Not level 4 because that boundary lives in a comment, not the description.