finish-feature
Finish a feature — run verification, create PR, retain learnings, and hand off cleanup to the terminal.
51
55%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/finish-feature/SKILL.mdQuality
Discovery
32%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 conveys a multi-step workflow for finishing features but lacks an explicit 'Use when...' clause, making it unclear when Claude should select this skill. The actions listed are somewhat specific but use vague terms like 'retain learnings' and 'hand off cleanup to the terminal' that don't clearly communicate concrete capabilities. Adding explicit trigger conditions and more natural user-facing keywords would significantly improve skill selection accuracy.
Suggestions
Add an explicit 'Use when...' clause, e.g., 'Use when the user says they are done with a feature, ready to submit a PR, or wants to wrap up and ship their changes.'
Expand trigger terms to include natural variations like 'pull request', 'submit PR', 'ship feature', 'done coding', 'wrap up', 'finish implementation'.
Clarify vague actions: specify what 'run verification' means (e.g., 'run tests and linting'), what 'retain learnings' means (e.g., 'save notes to a knowledge base'), and what 'hand off cleanup to the terminal' entails.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names several actions (run verification, create PR, retain learnings, hand off cleanup) but they remain somewhat vague — 'run verification' and 'retain learnings' lack specificity about what kind of verification or what learnings are retained. | 2 / 3 |
Completeness | Describes what the skill does (finish a feature workflow) but has no explicit 'Use when...' clause or equivalent trigger guidance, which per the rubric should cap completeness at 2, and the 'what' itself is only moderately clear, placing this at 1-2. Given the missing 'when' clause entirely, scoring at 1. | 1 / 3 |
Trigger Term Quality | Includes some relevant terms like 'PR', 'feature', and 'cleanup', but misses common natural variations users might say such as 'pull request', 'done with feature', 'ship it', 'merge', or 'finish up'. The phrase 'hand off cleanup to the terminal' is somewhat unusual jargon. | 2 / 3 |
Distinctiveness Conflict Risk | The combination of PR creation, verification, and learnings retention is somewhat distinctive, but 'finish a feature' is broad enough to overlap with general CI/CD skills, PR creation skills, or code review skills. | 2 / 3 |
Total | 7 / 12 Passed |
Implementation
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a strong, well-structured workflow skill with excellent actionability and clear validation checkpoints. The step-by-step process is logical with explicit stop conditions on failures and concrete commands throughout. Minor improvements could be made in conciseness (trimming some explanatory text) and potentially splitting the summary template into a referenced file for cleaner progressive disclosure.
Suggestions
Tighten prose in steps 3 and 6 — phrases like 'If any exist, surface them to the user' and 'Review the session and retain important discoveries' could be more concise since Claude understands these actions implicitly.
Consider extracting the summary template in step 7 into a referenced file or making it more compact inline to improve progressive disclosure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Generally efficient but has some unnecessary verbosity — the PR body guidance, the cleanup template, and some explanatory phrases could be tightened. The preamble section with ANSI escape codes and mark_chapter is appropriately specific. Some instructions like 'Confirm you are on a feature branch' are slightly redundant given the context. | 2 / 3 |
Actionability | Highly actionable with concrete, executable bash commands at every step. Git commands, gh CLI usage, devflow commands, and fallback alternatives are all copy-paste ready. The commit message format, PR body structure, and summary template are all specific and concrete. | 3 / 3 |
Workflow Clarity | Excellent multi-step workflow with clear sequencing and explicit validation checkpoints. Step 2 has a hard stop on failing checks ('Do NOT continue past this step'). Step 3 provides a diff self-test with clear remediation options. The workflow has a logical progression: assess → verify → review → commit → push → retain → summarize, with feedback loops for failures. | 3 / 3 |
Progressive Disclosure | The content is well-structured with clear sections (Preamble, Steps, Important), but it's a moderately long single file with no references to external documentation. The summary template and cleanup instructions could potentially be split out, though for a single-purpose skill this is acceptable. No bundle files are provided to offload detail. | 2 / 3 |
Total | 10 / 12 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 — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- AndreJorgeLopes/devflow
- Commit
b0b1bb6
- 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.