Conciseness

The body is mostly lean — command blocks, bullet lists, a 6-step description of a bundled script Claude cannot otherwise know about. Minor trimmable redundancy: the Authentication section states 'GITHUB_TOKEN environment variable provides automatic authentication' and then repeats 'If GITHUB_TOKEN is set, gh authenticates automatically for all API calls', and the 'Time-sensitive' guideline applies to the hardcoded 'gh version 2.87.2 (2026-02-20)' line, which is not in a deprecated/old-patterns section. This fits the score-4 anchor 'Efficient; minor instances of over-explanation that could be trimmed' rather than the score-5 'every token earns its place'.

Actionability

Most guidance is copy-paste ready: 'uv run .claude/skills/gh/scripts/setup_gh.py', a concrete CLI-options block, 'gh <command> -R <owner/repo>', and github_project_setup.py subcommands with real flags ('milestone create --title "v1.0" --due 2026-03-31'). The gap is the 'Common Commands' section, whose title promises actual gh command examples but contains only a dynamic '!...setup_gh.py --detect-only' invocation — the common pr/issue/run commands themselves are never shown inline. This places it at 'Mostly executable guidance; concrete code or commands with minor gaps' rather than the fully-executable score-5 anchor.

Workflow Clarity

The setup sequence is clear and ordered: detect gh missing → run setup script (with SHA256 verification and 401/403 auth fallback described) → verify with 'gh auth status' → apply the mandatory -R rule. Most checkpoints are present, but a post-install verification (e.g., 'gh --version') is only implicit in the script description, and the batch automation operations (label setup, issue import) have no stated validation/rollback guidance in SKILL.md. That matches score 4, 'Clear sequence with most checkpoints present; minor validation gaps' — not 3 (the sequence is explicit and validation exists for install) and not 5 (no explicit feedback loop for the batch operations).

Progressive Disclosure

The body is a ~130-line overview: core install/auth/-R guidance inline, with the four detail domains (labels, milestones, Projects V2, issue stories) split into one-level-deep references, each listed with a one-line description under 'Reference Files' ('[labels.md](./references/labels.md) — Label taxonomy...'). All six referenced paths were verified to exist on disk. This matches the score-5 anchor 'Clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation'.

Specificity

The description lists many concrete actions with no filler: 'Install and configure the GitHub CLI (gh)', 'auto-install script with SHA256 verification', 'GITHUB_TOKEN auth with anonymous fallback', 'required -R flag pattern', 'milestones via REST API', 'label taxonomy management'. This matches the anchor 'Lists multiple specific concrete actions; comprehensive coverage' — well beyond the score-4 anchor ('several specific actions; minor gaps'), as it covers install, auth, usage pattern, and four management domains.

Completeness

It explicitly answers what ('Install and configure the GitHub CLI... Provides auto-install script with SHA256 verification...') and when ('Use when gh command not found, shutil.which returns None, need GitHub API access...'). This is a textbook match for the score-5 anchor 'Clearly and explicitly answers both what AND when with concrete trigger phrases'; the explicit 'Use when' clause also avoids the cap noted in the guidelines.

Trigger Term Quality

Trigger phrases are exactly what a user/agent would encounter or say: 'gh command not found', 'shutil.which returns None', "repository operations fail with 'failed to determine base repo' error", 'need GitHub API access for issues or PRs or releases or workflow runs'. Both natural language and literal error strings are covered — comprehensive including synonyms and exact tool vocabulary, matching the score-5 anchor; score 4 ('a few natural terms missing') would understate the coverage.

Distinctiveness Conflict Risk

It carves a clear niche — gh setup and usage in proxied agent environments — with distinct triggers (gh-not-found, the 'failed to determine base repo' error, GITHUB_TOKEN, -R flag). The score-4 anchor ('minor overlap risk with closely related skills') is the nearest competitor, but the triggers are tool-specific and unlikely to fire a generic git or GitHub-API skill, so the score-5 'clear niche with distinct triggers' anchor fits best.