Conciseness

Mostly efficient — every example is domain-specific v1 syntax Claude would not know, with no filler explaining general concepts — but the v0→v1 contrast rules are stated three to four times ("Use `script:` field in run steps, never `command:`" appears in Instructions, the Key Differences table, Performance Notes, and Troubleshooting) and the "consult references/v1-spec-schema.md" directive appears in three places. Not a 4 because this cross-section repetition is more than minor trimming; not a 2 because there is no padding with concepts Claude already knows.

Actionability

Fully executable guidance throughout: copy-paste-ready YAML for every step type (run, run-test, action, template, approval, background), complete CI and CD pipeline examples, and an exact MCP invocation ("resource_type: \"pipeline\"... body: { yamlPipeline: \"<full v1 pipeline YAML string>\" }") with concrete error-code fixes ("DUPLICATE_IDENTIFIER — Pipeline exists; use harness_update"). This matches the anchor-5 'copy-paste ready, covers the common cases' example.

Workflow Clarity

A clear six-step sequence (confirm v1 format → clarify requirements → consult spec → select native actions → generate YAML → optionally create via MCP) with a verification checkpoint before the risky MCP operation ("Verify the project exists" with harness_list) and error-recovery guidance in Troubleshooting. Not a 5 because the validation directive ("Validate all expressions use ${{ }} syntax before presenting") sits in Performance Notes rather than as an explicit step in the main workflow, leaving a minor checkpoint gap.

Progressive Disclosure

Structure is good: two real, substantive one-level-deep reference files (references/v1-spec-schema.md and references/native-actions.md) are clearly signaled at point of use ("Consult the spec reference - Use `references/v1-spec-schema.md` for the complete v1 schema... the full action catalog"), with the body holding key patterns and the bulk catalogs in the references. Not a 5 because the 465-line body duplicates catalog content (e.g., the action/template example gallery restated in Instructions and Step Types) that could be pushed to the references.

Specificity

The description lists multiple specific concrete actions with comprehensive coverage: "Generate Harness v1 simplified Pipeline YAML using the new concise syntax with flat structure, ${{ }} expressions, script field, and action steps" plus enumerated capabilities ("CI stages (run, run-test, background)", "CD stages... for K8s, Helm, ECS", "approval (stage-level and inline)", "parallel execution, matrix/for/while strategies, caching, volumes, and templates"). Not a 4 because the gaps are negligible — the entire capability surface is named with concrete syntax terms rather than generic verbs.

Completeness

It clearly and explicitly answers both what ("Generate Harness v1 simplified Pipeline YAML..." with enumerated features) and when ("Use when asked for a v1 pipeline, simplified pipeline, new pipeline format, or when user specifically requests v1 syntax"), and adds a negative trigger ("Do NOT use for v0/standard pipelines"). This matches the anchor-5 pattern of concrete trigger phrases attached to both a what and a when, so a 4's 'could be more explicit' caveat does not apply.

Trigger Term Quality

Keyword coverage is good: it enumerates natural trigger phrases ("v1 pipeline, simplified pipeline, new pipeline format, create v1, modern pipeline syntax") that a user would plausibly say, and the "Use when asked for a v1 pipeline..." clause reinforces them. Not a 5 because a few natural variants are missing (e.g., "Harness pipeline", "concise pipeline", "flat pipeline syntax"); not a 3 because the explicit phrase list covers the common phrasings well rather than leaving only domain jargon.

Distinctiveness Conflict Risk

It carves a clear niche with minimal conflict risk by explicitly fencing off the sibling skill: "Do NOT use for v0/standard pipelines (use create-pipeline)" alongside v1-specific triggers. This is exactly the anchor-5 'clear niche with distinct triggers' pattern; a 4's 'minor overlap risk' is not evident since the v0/v1 boundary is stated explicitly.