Conciseness

The body consistently adds only CLI-specific knowledge Claude would not have (auto-prefix quirk, ghost thresholds, grep patterns) and never pads with known concepts — but the same critical rules are repeated verbatim-ish many times: ghost accumulation appears in ~5 places (Mental Model, "Ghost accumulation is silent", the full §Ghost Discipline section, the pitfall table, and M-2), the ≥5cm/≥15° spatial-variety rule in 3 places, and the `$9/mo` single-quoting rule in 3. Fits the 3 anchor ("mostly efficient… could be tightened"); not 2 because nothing explains concepts Claude already knows, and not 4 because the duplication is extensive enough that the ~470-line file could shrink by a third without losing information.

Actionability

Fully executable, copy-paste-ready guidance throughout: four complete recipe blocks with exact `officecli add/set` commands and props, a runnable Python helper-invocation script, the M-4 `raw-set` speed example with readback, jq/grep one-liners, and four complete gate scripts with REJECT criteria. Specific examples cover the common morph cases (single element, coordinated actors, long arc, hybrid fade).

Workflow Clarity

Clear sequence (read pptx v2 → plan the §Morph Pair Planning table BEFORE coding → build via recipes → run gates → deliver) with explicit validation checkpoints and feedback loops for a batch operation: Gates 1–5a inherited, a morph-specific Gate 2 grep addendum, and Gate 5b-morph-1..4 proof loops, capped by "If any Gate prints REJECT, fix and re-run — never deliver with a known-open gate". Matches the 5 anchor (explicit validation, error-recovery loops, checklists).

Progressive Disclosure

References are one level deep, well-signaled, and each is annotated with when to read it (References section plus inline "→ see pptx v2 §X" pointers covering decision-rules.md, pptx-design.md, morph-helpers.py/.sh, styles/INDEX.md). Falls short of 5 because the SKILL.md is a ~470-line full manual rather than an overview — recipes, gate scripts, ghost-discipline detail, and the style-library workflow are all inline where some belong in reference files — and no bundle files are present in this skill directory, so the `reference/` paths cannot be verified to exist.

Specificity

Concrete capabilities are explicit — "wants a .pptx with smooth cross-slide animation", "shapes that grow / move / rotate as the slide advances", "Output is a single .pptx" — but they are framed as output characteristics rather than a comprehensive list of executable actions. Not the 5 anchor because it never enumerates what the skill actually does (build slides, bind shape names, choreograph transitions); not 3 because multiple specific, non-generic capabilities are stated.

Completeness

Both what ("a .pptx with smooth cross-slide animation… Output is a single .pptx", "scene layer on top of officecli-pptx") and when ("Use this skill when…", explicit trigger list, plus negative guidance "DO NOT invoke for a generic deck, pitch deck, or board review without cross-slide motion") are clearly and explicitly answered with concrete trigger phrases.

Trigger Term Quality

Eight explicit natural trigger phrases ("'morph', 'morph transition', 'smooth transition', 'continuous animation across slides', 'Keynote-style transition', 'animated slide sequence', 'shape continuity across slides'") plus synonyms ("PowerPoint Morph transitions", "Keynote-style continuous motion") and the file extension ".pptx". Comprehensive coverage matching the 5 anchor.

Distinctiveness Conflict Risk

Clear niche (cross-slide morph motion only) with distinct triggers and explicit conflict-resolution routing: "route those to officecli-pptx base or officecli-pitch-deck". Minimal overlap risk with the sibling skills it names.