Conciseness

The body is ~8 lines with zero padding: "新任务读 `../avatar/SKILL.md` 并用 `avatar generate`" and the bare resume command carry all the information, assuming Claude's competence with no explanations of known concepts. This squarely matches anchor 5 ('lean and efficient; assumes Claude's competence; every token earns its place').

Actionability

One concrete, copy-pasteable command is present ("liveportrait resume --job /absolute/render/presenter.liveportrait.json"), but key details are missing: the body never explains how the `liveportrait` command maps to the bundled `scripts/liveportrait.py`, the referenced `../avatar/SKILL.md` does not exist in the bundle, and the mandated acceptance steps ("原声哈希、时长验收") have no executable commands. This matches anchor 3 ('some concrete guidance but incomplete; missing key details') rather than anchor 4's 'minor gaps'.

Workflow Clarity

A rough sequence exists (run resume, preserve fields, gate downstream composition on acceptance), and validation requirements are stated ("timing_pass=false 的记录不得手改为通过", "验收通过后才能参与后续合成"), but the checkpoints are implicit — no commands for running the hash/duration acceptance and no recovery path when resume fails or timing_pass=false. This fits anchor 3 ('sequence present but checkpoints missing or implicit'), not anchor 4 which requires most checkpoints with only minor gaps.

Progressive Disclosure

Against the actual bundle structure (SKILL.md + scripts/liveportrait.py), the body is short and sectioned but never points to `scripts/liveportrait.py` that implements the `liveportrait` command, and its only external reference, `../avatar/SKILL.md`, is not a real file in this bundle. That matches anchor 3 ('some structure; references present but not clearly signaled') rather than anchor 4's 'most content appropriately placed; references mostly clear'.

Specificity

The description names the domain and 1-2 concrete actions — "旧数字人任务续查兼容入口" (compatibility entry for resuming legacy digital-human tasks) and routing new tasks to "包内 avatar 工具" — but lists no further concrete capabilities (e.g., checking job status, verifying timing/hash acceptance). This matches anchor 3 ('names domain and 1-2 concrete actions, not comprehensive') and falls short of anchor 4, which expects several specific actions.

Completeness

The 'what' is clear (resume legacy expert-video digital-human tasks), and 'when' is explicitly stated through the routing rule "新任务使用包内 avatar 工具" — use this entry only for old-task resumption, not new generation. Both are present with the 'when' expressible more directly (no 'Use when...' phrasing or concrete trigger phrases), matching anchor 4 rather than anchor 5; it is above anchor 3 because the when is explicitly stated, not merely implied.

Trigger Term Quality

It includes some relevant keywords a user might say — "expert-video", "旧数字人任务", "续查", "avatar" — but misses common variations and synonyms such as "数字人视频", "口播", "liveportrait", or job-file extensions (.liveportrait.json). This is anchor 3 ('some relevant keywords but missing common variations or synonyms'), not anchor 4's 'good keyword coverage with only a few natural terms missing'.

Distinctiveness Conflict Risk

The description carves out a clear niche (legacy-task resumption only) and explicitly disambiguates against the closest sibling skill ("新任务使用包内 avatar 工具"), which is anchor 4's 'mostly distinct; minor overlap risk with closely related skills'. It stops short of anchor 5 because the shared "数字人" domain still overlaps with the avatar skill and no distinct trigger phrases separate them.