Conciseness

The body spends its tokens on non-obvious domain knowledge (why the host's 'idle' label is insufficient, why conversation memory is not a source) and gives git commands without library-primer padding, so nothing explains concepts Claude already knows. A few rule rationales could still be tightened, which keeps it at anchor 4 rather than 5, but there is no real padding that would drop it to 3.

Actionability

The instructions are copy-paste executable throughout: `git show <branch>:<task file path>`, `git rev-list --count <main>..<branch>`, `git branch --merged <main>`, `git rev-parse --verify --quiet refs/remotes/origin/<branch>`, `git -C <worktree> status --porcelain`, with explicit fallbacks ("or 'no remote' when `git remote` prints nothing") and an enumerated per-child board row. Placeholders like <branch> are inherent to the task, not pseudocode, matching the level-5 anchor.

Workflow Clarity

Five explicitly sequenced steps with an explicit verification checklist and concrete error handling for every failure path ("No `## Swarm` section → say in one line… and stop", "A missing or unreadable child task file is a row `state unknown`", "No such tool → say in one line that sessions cannot be listed"). The skill is read-only reporting, so the destructive/batch cap does not apply; this matches anchor 5's checklist plus error-recovery rather than anchor 4's 'most checkpoints'.

Progressive Disclosure

The skill is a self-contained ~110-line document with clean sections (Overview, Context, Rules & Constraints, Instructions, Verification) and no content that belongs in a separate file inlined — no references/scripts/assets exist and none are needed. It scores anchor 4 rather than 5 because it exceeds the under-50-line simple-skill case and offers no reference structure to reward, only well-organized inline sections.

Specificity

"builds one board from the children's task files, branches and the host's session list, with the operator's next steps" names the artifact, three concrete data sources, and the output. It stops short of the level-5 anchor's multiple distinct actions (it is one composite action), but exceeds level 3's '1-2 concrete actions'.

Completeness

"Use when a manager session is asked for the state of its child sessions" is an explicit 'when' with a concrete trigger phrase, and "builds one board from the children's task files, branches and the host's session list, with the operator's next steps" is a clear, concrete 'what'. Both are explicitly answered, matching the level-5 anchor; level 4 would require the 'when' to be less explicit than it is.

Trigger Term Quality

"asked for the state of its child sessions" and "manager session" are natural phrases an operator would say when they need this skill, matching anchor 4's 'good keyword coverage; a few natural terms missing'. It is not level 5 because common variations such as 'status' or 'progress' of the children are absent.

Distinctiveness Conflict Risk

"Not for one session's own report (session-status)" explicitly disambiguates the closest sibling skill, and the manager/children-sessions niche is narrow with distinct triggers. This matches anchor 5's 'clear niche with distinct triggers; minimal conflict risk'.