Conciseness

The body is imperative and dense with almost no explanation of concepts Claude already knows, but contains some redundancy: 'Don't overwrite user-written assessments with API boilerplate' appears in both Validation Rules and Anti-Patterns, and the back-linking rule is stated in the Contract, a blockquote, and Step 7's note. Anchor 4: efficient with minor instances that could be trimmed; not 5 because of the duplicated rules.

Actionability

Concrete commands appear throughout ('gbrain search "name"', 'gbrain lint', 'gbrain wanted', 'gbrain sweep --once', 'gbrain timeline-add'), alongside copy-paste-ready person and company page templates and a JSON example for put_raw_data. Anchor 5: fully executable, copy-paste-ready guidance covering the common create and update cases.

Workflow Clarity

The 7-step protocol is clearly sequenced with CREATE/UPDATE branches, and validation is explicit: notability gate, 'gbrain lint' flags empty sections, 'Test on 3-5 entities first. Read actual output. Check quality.' before bulk runs, 'Commit every 5-10 entities', and a 'When it fails' section with error-specific recovery (revision_conflict re-read/merge/save, rate_limited back-off, auto_links.errors slug fix plus add_link). Anchor 5: clear sequence, explicit validation checkpoints, and feedback loops for batch operations.

Progressive Disclosure

No bundle files ship with the skill; the cross-skill references ('skills/_brain-filing-rules.md', 'skills/conventions/quality.md', 'skills/brain-taxonomist/SKILL.md', '../conventions/agent-operator-protocol.md') are one level deep and clearly signaled via blockquote callouts. The two full page templates (~90 lines) are inlined rather than split into reference files, which fits anchor 4's 'minor organization gaps' rather than 5's clean overview-plus-references split.

Specificity

Names the domain ('brain pages', 'person/company pages') and several concrete actions ('Creates and updates', 'compiled truth, timeline, and cross-links'), but coverage is not comprehensive — external research, validation, and reporting go unmentioned. Anchor 4 fits: several specific actions with minor gaps; not 5 because key capabilities are unlisted, not 3 because more than 1-2 actions are concrete.

Completeness

Both what ('Creates and updates person/company pages with compiled truth, timeline, and cross-links') and when ('Use when a new entity is mentioned or an existing page needs updating') are explicitly stated. Anchor 4 fits because the when clause could name more concrete trigger situations; not 5 since trigger phrases are not comprehensive, not 3 since when is explicit rather than implied.

Trigger Term Quality

Natural terms like 'enrich', 'person/company pages', 'new entity is mentioned', and 'existing page needs updating' give good keyword coverage, but common user phrasings such as 'who is this person' or 'look up this company' appear only in the triggers list, not in the description itself. Anchor 4 rather than 5 (variations missing) or 3 (coverage is better than 'some relevant keywords').

Distinctiveness Conflict Risk

The personal-brain enrichment niche ('brain pages', 'compiled truth', 'cross-links') is clearly distinguishable from generic document or search skills, with only minor overlap risk from 'enrich' and entity-lookup phrasing that related knowledge-base skills might share. Anchor 4: mostly distinct with minor overlap risk.