Conciseness

The body is efficient overall — a tool table, concrete parameterized calls, and tight error tables with no explanations of concepts Claude already knows. However, the 'Performance Notes' section is largely redundant: 'List existing roles and resource groups before creating new ones to avoid duplication' and 'Verify role permissions match the principle of least privilege' repeat Best Practices and Error Handling content, fitting the level-4 anchor 'Efficient; minor instances of over-explanation that could be trimmed' rather than the lean level-5 anchor.

Actionability

Concrete executable MCP tool calls with parameters (e.g., the full 'harness_create' body with identifier, permissions list, and the identifier regex), realistic examples, and a cause/solution error table fit 'Mostly executable guidance; concrete code or commands with minor gaps'. Not a 5 because the core 'grant access' use case is never shown: there is no 'harness_create' example for 'role_assignment', even though the description promises 'assign roles' and 'grant access'.

Workflow Clarity

Steps 1-6 are clearly sequenced with early pre-checks (list existing roles, check current assignments), and Troubleshooting provides feedback loops. However, the skill includes destructive operations ('harness_delete' on roles) with no explicit validation or verification checkpoint before deletion (e.g., checking assignments that reference the role or confirming with the user), which per the destructive-operations guideline caps workflow clarity at 3: 'Steps listed but validation gaps; checkpoints missing or implicit'.

Progressive Disclosure

SKILL.md is a well-organized overview (tools, instructions, examples, errors) that offloads the bulk lookup data to a single one-level-deep reference, clearly signaled: 'For built-in roles (account/org/project/module), resource groups, common permissions, and role assignment structure, consult references/builtin-roles.md' — and that file exists in the bundle. This matches the anchor 'Clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation'.

Specificity

The description lists multiple concrete actions across the full RBAC domain: 'List, create, update, and delete custom roles. View role assignments and permissions for users, groups, and service accounts' alongside 'Manage Harness RBAC roles, role assignments, permissions, and resource groups via MCP v2 tools'. This matches the anchor 'Lists multiple specific concrete actions; comprehensive coverage' rather than the level-4 anchor, which requires minor gaps in coverage.

Completeness

It explicitly answers both 'what' (first three sentences of concrete capabilities) and 'when' ('Use when asked to manage access control, assign roles, check permissions, create custom roles, review RBAC configuration, onboard users, or audit access') with a dedicated concrete trigger-phrase list. This matches the anchor 'Clearly and explicitly answers both what AND when with concrete trigger phrases'.

Trigger Term Quality

Trigger phrases include natural synonyms and variations users would actually say: 'manage roles, RBAC, role assignment, user permissions, access control, custom role, resource group, who has access, grant access, revoke access', plus 'Use when asked to... assign roles, check permissions... onboard users, or audit access'. This matches the anchor 'Comprehensive coverage of natural terms including synonyms'.

Distinctiveness Conflict Risk

The 'Harness RBAC... via MCP v2 tools' framing gives it a clear niche with distinct access-control triggers, fitting the level-5 anchor. However, 'resource group' and 'onboard users' as trigger phrases carry minor overlap risk with adjacent Harness admin skills (e.g., a resource-group or user-management skill), which fits the level-4 anchor 'Mostly distinct; minor overlap risk with closely related skills' slightly better than 5.