Conciseness

The ~10-line body is lean with no padding or explanation of concepts Claude already knows — every line ("Set the old wording aside", "Push the derived statement shorter, pass by pass") carries instruction. Only a faint echo of the shortest-covers-everything preference between paragraphs 2 and 3 keeps it from being pure anchor-5 material, but it still fits "every token earns its place" better than anchor 4's "minor instances of over-explanation."

Actionability

Concrete, directive guidance throughout: "derive the shortest statement that covers all of it", "Try a tree or a table as a candidate... it wins only by being shorter", "back off by the minimum that recovers it", plus a defined output ("Hand back the reduced statement alone"). It falls short of anchor 5 because there is no worked example of an input text and its reduced output, which the top anchor requires ("specific examples cover the common cases").

Workflow Clarity

The sequence is clear (confirm closed set → set old wording aside and derive → shorten pass by pass → hand back the result alone → escalate to **widen** otherwise) and it contains an explicit validate-and-recover loop: "When a pass makes it stop covering something, it went too far — back off by the minimum that recovers it." Not anchor 4 because no checkpoints are missing; the per-pass coverage check, back-off rule, floor ("a clause that survives every pass is load-bearing"), and termination condition are all explicit.

Progressive Disclosure

The skill is under 50 lines with no detail that belongs in a separate file, so per the rubric's simple-skill guidance well-organized content alone merits a 5. Paragraphs are logically ordered (scope, method, moved-target case, iteration loop, output, escalation) and the single external pointer ("use **widen** skill") is one level deep and clearly signaled.

Specificity

"Re-derive something written into the shortest version that still covers everything" names one concrete action with a precise success criterion, matching anchor 3 ("1-2 concrete actions, but not comprehensive") rather than anchor 4's "several specific actions" — there is no breadth of named capabilities or artifact types.

Completeness

"Re-derive something written into the shortest version that still covers everything" explicitly answers what, and "Use when something already written should get shorter and nothing it covers can be lost" explicitly answers when with a concrete trigger condition — matching anchor 5 rather than 4, where the "when" would be weaker or less specific.

Trigger Term Quality

"something already written should get shorter" is a natural phrase a user might say, but the description misses common variations and synonyms ("condense", "shorten", "tighten", "trim"), fitting anchor 3 ("some relevant keywords but missing common variations or synonyms") rather than anchor 4's "good keyword coverage, a few natural terms missing".

Distinctiveness Conflict Risk

The lossless constraint ("nothing it covers can be lost") carves out a niche distinct from summarizing or rewriting, so it is mostly distinct with minor overlap risk against closely related editing/summary skills — anchor 4. Not anchor 5 because the single-phrased trigger leaves some conflict room with generic shortening/editing skills.