Conciseness

The body is compact bullet lists with no explanations of concepts Claude already knows — every section is short and directive. Not 5 because the Philosophy bullets ("Produce reproducible architecture artifacts before making architecture claims", "Start from live evidence and local patterns") and parts of Anti-Patterns largely duplicate Workflow/Avoid and could be trimmed; not 3 since there is no padded or unnecessary conceptual explanation.

Actionability

"Run deterministic diagram/context-pack commands" gives no command names, syntax, or invocation examples anywhere in the body — the core executable detail is missing, and the Examples section lists user prompts ("Refresh the architecture context pack for this repo") rather than runnable commands. There is concrete guidance ("Report exact failed command and artifact path", "Stop on the first broken artifact gate") and a pointer to references/contract.yaml, but that contract also contains no commands, so guidance remains incomplete rather than minimal — matching the 'some concrete guidance but missing key details' anchor, not the 2 anchor's high-level-hints-only level.

Workflow Clarity

The Workflow section gives a clear ordered sequence with an explicit validation checkpoint ("Validate manifest completeness and placeholder status") and a fail-fast feedback loop ("Stop on the first broken artifact gate"; "stop at first failed gate; do not proceed until it is fixed and rerun"), so batch operations are gated. Not 5 because the steps stay abstract — no concrete commands, and the validation mechanics are delegated elsewhere; comfortably above 3 since checkpoints and error-recovery loops are explicit rather than implicit.

Progressive Disclosure

The Progressive Disclosure section clearly signals one-level-deep references, and references/contract.yaml, references/evals.yaml, and references/task-profile.json all exist in the bundle. Not 5 because "Infrastructure/references/deferred-skill-context/agent-ops-diagram-cli/" does not exist in the bundle — a dangling navigation path; not 3 since structure and signaling are otherwise genuinely good with the body appropriately kept lean.

Specificity

Lists three concrete actions — "Generate, validate, and refresh" — against a named tool/domain ("@brainwav/diagram architecture artifacts"), with artifact types ("repo diagrams, context packs") surfacing in the trigger clause. Falls short of 5 because coverage of what is actually produced is thinner than the comprehensive multi-action anchor; above 3 because it exceeds the 1-2-action anchor.

Completeness

Explicitly answers both questions: the 'what' is "Generate, validate, and refresh @brainwav/diagram architecture artifacts" and the 'when' is the explicit clause "when repo diagrams, context packs, PR impact, or CI drift evidence is needed" with concrete triggers. This matches the anchor for a clear 'what' plus explicit 'when' with concrete trigger phrases, so it does not cap at 3 or sit at 4.

Trigger Term Quality

"repo diagrams", "context packs", "PR impact" are terms users would plausibly say, giving good keyword coverage. Not 5 because "CI drift evidence" and "@brainwav/diagram" are tool-specific jargon and natural synonyms like "architecture map" or "onboarding diagram" are absent; not 3 since several genuinely natural trigger phrases are present, not just generic keywords.

Distinctiveness Conflict Risk

The niche is distinct — architecture evidence artifacts tied to a specific @brainwav/diagram toolchain with triggers like "context packs" and "CI drift evidence". Minor overlap risk remains because "repo diagrams ... needed" could catch generic diagram-drawing requests; not 5 for that residual overlap, not 3 because the tool anchoring keeps it well above the merely-somewhat-specific anchor.