finish-feature

Finish a feature — run verification, an adversarial review gate (re-grounded on the original design doc + ticket ACs, not the plan) and a deferral-closure gate, then create the PR, retain learnings, and hand off cleanup to the terminal.

Quality

—

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Advisory

Suggest reviewing before use

Quality

Content

77%

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

The content is highly actionable with a well-sequenced, validation-checkpointed pipeline and executable commands throughout. The main weakness is conciseness — the rationale blockquotes add tokens that assume less of Claude than the rest of the skill does — and the absence of any bundle/reference split for a skill this long.

Suggestions

Trim the Step 4 'Why this step exists' blockquote and inline rationale to one or two lines, moving the full justification to a reference file if needed.

Consider extracting the long gating rules (e.g. the full deferral-closure and review-gate criteria) into a referenced doc to improve progressive disclosure for a 190-line skill.

Move the deterministic-predicate contract explanation out of the body into a short pointer, since it largely re-explains a shared lib Claude can source directly.

Dimension	Reasoning	Score
Conciseness	The body is mostly lean — executable bash, regex, and pinned formats assume Claude's competence — but the Step 4 'Why this step exists' blockquote and several parenthetical justifications are explanatory prose that could be tightened without losing the instruction.	2 / 3
Actionability	Every step ships executable commands or exact formats: branch-guard predicate, `devflow check`, merge-base diff, a conventional-commit regex, pinned PR headings, and `count_commits` fallbacks — copy-paste ready with no pseudocode.	3 / 3
Workflow Clarity	A clearly sequenced 9-step pipeline with explicit validation checkpoints and feedback loops — hard-stop branch guard, 'If checks fail, stop', 'Fail loud' empty-diff check, severity-gated findings, and the deferral-closure gate — matching the anchor for clear sequence with explicit validation.	3 / 3
Progressive Disclosure	No bundle files exist, and the ~190-line SKILL.md is a single well-sectioned document that references external paths (~/.claude/lib/determinism/, MCP) not in a bundle; it is organized but monolithic rather than split into one-level-deep references.	2 / 3
	Total	10 / 12 Passed

Description

60%

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

The description is specific and third-person, listing a concrete sequence of completion actions, but it lacks an explicit 'Use when' trigger that tells Claude when to invoke it. Adding natural trigger phrases would raise completeness and trigger-term quality.

Suggestions

Append a 'Use when finishing a feature / completing work and opening the PR / finalizing a ticket before cleanup' clause so Claude knows when to trigger it.

Add common user-facing variations like 'wrap up', 'finalize', 'open the PR', 'feature complete' to broaden trigger coverage.

Lead with the trigger condition ('Use when a feature is implemented and ready to finalize') before the action list to sharpen distinctiveness from generic PR skills.

Dimension	Reasoning	Score
Specificity	Names multiple concrete actions — 'run verification', 'an adversarial review gate', 'a deferral-closure gate', 'create the PR', 'retain learnings', and 'hand off cleanup' — matching the 'lists multiple specific concrete actions' anchor.	3 / 3
Completeness	It clearly states what the skill does but provides no explicit 'when should Claude use it' trigger; the judging guidelines state a missing 'Use when...' clause should cap completeness at 2.	2 / 3
Trigger Term Quality	Only 'finish a feature' is a natural phrase a user might say; there are no 'Use when...' clauses or common keyword variations ('finish my work', 'complete this feature', 'open the PR'), so coverage is partial rather than comprehensive.	2 / 3
Distinctiveness Conflict Risk	The named gates give it a recognizable niche, but 'finish a feature / create the PR' could still overlap with generic PR-creation or wrap-up skills, so it is not yet a clearly conflict-free trigger.	2 / 3
	Total	9 / 12 Passed

Validation

100%

Warnings & errors only

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: AndreJorgeLopes/devflow
Commit: 9556c94

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

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.