Conciseness

The body is command-first and dense with executable blocks, and tool-specific facts (dev-url scratch-DB semantics, hash tamper guard) earn their tokens. Minor trimmable padding remains: the Flyway/Liquibase value-add pitch in Step 6, the repeated lint-gate instruction in 'How to use' and Step 5, and re-explained NOT NULL default logic in the worked example — anchor 4 ("Efficient; minor instances of over-explanation that could be trimmed") rather than 5.

Actionability

Nearly every step ships a copy-paste command with full flags (schema apply, migrate diff, migrate apply, CI YAML, worked example). Two minor gaps keep it from anchor 5: Step 6 uses a placeholder `--dev-url "..."` instead of a concrete value, and `atlas migrate hash` is described in Step 7 but never shown with its invocation.

Workflow Clarity

Steps 1-9 are explicitly ordered with a hard validation gate: "run Step 6 lint first and assert it reports no destructive / data-loss findings; if it flags one, fix the migration ... and re-run lint until clean. Only then apply" — a full validate/fix/retry feedback loop for destructive database operations, plus error recovery for the hash-mismatch failure in Step 7 and the worked example. This matches anchor 5 (explicit validation steps, feedback loops) and clears the destructive-operation cap.

Progressive Disclosure

The SKILL.md body is a workflow overview; HCL syntax detail and the anti-patterns/limitations catalogue are split into references/hcl-schema.md and references/atlas-caveats.md, both real files, clearly signaled inline with links and a summary of what each contains, and one level deep. This matches the anchor-5 pattern of a concise overview with well-signaled one-level-deep references.

Specificity

Lists five concrete actions each bound to its exact command: "`atlas schema apply` for desired-state apply OR `atlas migrate diff` to generate versioned migrations", "`atlas migrate apply` to deploy", "`atlas migrate lint` to flag destructive / locking / data-loss patterns", "`atlas migrate hash` to detect tampering", plus the supported-database list. This matches the anchor "Lists multiple specific concrete actions; comprehensive coverage" and third-person voice ("Authors and runs") is used correctly; there is no vagueness to push it below.

Completeness

Both what and when are present: the what is explicit and command-level, and "Use when the user wants Terraform-style declarative DB schema management or modern SQL-first migration linting beyond Flyway / Liquibase" is an explicit when-clause. It stays at 4 rather than 5 because the when-clause covers only the declarative-management and linting motivations, not the equally common "apply/versioned migrations to a database" trigger.

Trigger Term Quality

Good natural-term coverage: "database schema migrations", "migrations", "deploy", "lint", "Terraform-style declarative", "Flyway / Liquibase". A few natural phrasings users would say are missing (e.g., "schema change", "DDL", "migration files", ".sql"), so it fits anchor 4 ("Good keyword coverage; a few natural terms missing") rather than anchor 5's comprehensive synonym/extension coverage.

Distinctiveness Conflict Risk

Atlas is named throughout with a clear niche (declarative/versioned schema migrations with lint and tamper detection), but "Terraform-style" and the bare "Flyway / Liquibase" mentions could pull this skill in when a user asks about those tools or a sister flyway-migrations/liquibase-migrations skill exists. Mostly distinct with minor overlap risk — anchor 4.