CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

build-feature

Use whenever building or changing a feature in Skylark — the loop from first failing test to an open, mergeable PR. Covers red-green TDD, making `npm run check` pass, committing on a branch, pushing, and opening a PR. Once the PR is open, hand off to the **babysit-pr** skill (or a dedicated babysitter agent) to shepherd it through CI, reviews, and the merge — a feature isn't shipped until that's done too.

75

Quality

94%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

92%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is a lean, well-sequenced workflow with concrete commands and explicit validation checkpoints, delegating adjacent concerns (service scaffolding, PR babysitting) to sibling skills. Its single weakness is the unspecified issue-CLI invocation in the handoff step.

DimensionReasoningScore

Conciseness

The body assumes Claude's competence — it never explains TDD, git, or what a PR is — and every sentence carries project-specific information (the commit-gate/landing-gate behavior, PGlite-based tests, the coverage/mutation commands). It is lean with no padded sections, matching the 5 anchor.

5 / 5

Actionability

Most steps are executable with concrete commands (`npm run check`, `npm run coverage:diff`, `npm run mutate:diff`, `git push -u origin <branch>`, `gh pr create --fill`), but the project-specific issue-CLI handoff in step 5 ('Hand the baton to it through the issue CLI') gives no command or example. That minor gap fits the 4 anchor rather than the fully copy-paste-ready 5.

4 / 5

Workflow Clarity

The five numbered sections give a clear sequence with explicit validation checkpoints: watch the failing test fail, run check 'until it's clean', confirm coverage and mutation scores, and a feedback loop for failure ('if it fails it blocks the commit and hands you the errors — fix them, don't work around the gate'). This matches the 5 anchor's explicit validation and error-recovery loop.

5 / 5

Progressive Disclosure

The ~55-line body keeps everything appropriately inline at this size, is organized into clearly headed stages, and delegates detail that belongs elsewhere via well-signaled pointers to the create-service and babysit-pr skills. No bundle files exist and none are needed; navigation is easy, matching the 5 anchor.

5 / 5

Total

19

/

20

Passed

Description

92%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: it states a concrete, multi-step capability set, opens with an explicit 'Use whenever' trigger clause, and cleanly delimits its scope against the sibling babysit-pr skill. The only weakness is moderate synonym coverage in its trigger phrasing.

DimensionReasoningScore

Specificity

The description enumerates multiple concrete actions — 'red-green TDD, making `npm run check` pass, committing on a branch, pushing, and opening a PR' plus shepherding 'through CI, reviews, and the merge' — comprehensively covering the feature loop. Coverage is complete rather than having minor gaps, so it matches the 5 anchor rather than 4.

5 / 5

Completeness

It explicitly answers both questions: 'what' via the enumerated loop 'from first failing test to an open, mergeable PR' and 'when' via the explicit trigger clause 'Use whenever building or changing a feature in Skylark'. Both are concrete and explicit, matching the 5 anchor.

5 / 5

Trigger Term Quality

'Use whenever building or changing a feature' is a natural phrase users would say, supported by terms like test, commit, push, and PR. Common synonyms and variations ('implement', 'add X', 'open a PR') are missing, which fits the 4 anchor (good coverage, a few natural terms missing) rather than the comprehensive 5.

4 / 5

Distinctiveness Conflict Risk

The skill is scoped to a clear niche (the build-to-open-PR loop in Skylark) and explicitly disambiguates its closest neighbor: 'Once the PR is open, hand off to the babysit-pr skill'. Because the boundary with the overlapping sibling skill is explicitly drawn, conflict risk is minimal, matching the 5 anchor.

5 / 5

Total

19

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

drufball/skylark

Path

plugins/skylark-builder/skills/build-feature/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.