pantheon-ai/pr-decomposition
Splits a large feature branch into smaller, focused pull requests using stacked branches and cherry-pick. Groups commits by concern (infrastructure, application logic, tests, housekeeping), proposes descriptive branch names for user approval, creates stacked branches, and generates What/Why MR titles and descriptions.
83
83%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that clearly articulates specific capabilities (grouping commits, creating stacked branches, generating MR descriptions) and provides explicit trigger guidance with natural user language. It occupies a distinct niche that minimizes conflict risk with other skills, and uses proper third-person voice throughout.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: splits feature branches, groups commits by concern (infrastructure, application logic, tests, housekeeping), proposes descriptive branch names, creates stacked branches, generates What/Why MR titles and descriptions. Very detailed and actionable. | 3 / 3 |
Completeness | Clearly answers both 'what' (splits branches, groups commits, proposes branch names, creates stacked branches, generates MR descriptions) and 'when' with an explicit 'Use when...' clause listing four specific trigger scenarios. | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms users would actually say: 'PR is too big', 'split a PR', 'decompose a branch', 'break work into reviewable chunks', plus domain terms like 'stacked branches', 'cherry-pick', 'pull requests'. Good coverage of natural language variations. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche — splitting large PRs into stacked branches using cherry-pick is a very specific workflow. The triggers ('PR is too big', 'split a PR', 'decompose a branch') are unlikely to conflict with general git or code review skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
57%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured skill with excellent progressive disclosure and clear anti-patterns. Its main weaknesses are that the core workflow is too summarized inline (delegating actionability to a reference file) and the mindset/context sections explain things Claude already understands. The anti-patterns section, while valuable, partially duplicates content that's also in a referenced file.
Suggestions
Inline the key workflow steps with actual git commands (cherry-pick sequence, branch creation, target setting) rather than just a one-line summary pointing to the reference file — the skill should be executable without loading references for the common case.
Trim the 'Mindset' section to 1-2 sentences or remove it entirely — Claude understands why large PRs are problematic and doesn't need the explanation of reviewer cognitive load.
Add an explicit validation checkpoint after cherry-picking (e.g., 'After cherry-picking each group, run the test suite and verify compilation before proceeding to the next branch') rather than only mentioning it as an anti-pattern to avoid.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient but includes some unnecessary content. The 'Mindset' section explains concepts Claude likely understands (why large PRs are bad for reviewers). The 'When to use' and 'When not to use' sections add value but could be tighter. The anti-patterns section is somewhat verbose with repeated WHY/Consequence patterns that overlap with the referenced anti-patterns file. | 2 / 3 |
Actionability | The skill provides some concrete guidance (git log/diff-tree commands, branch naming examples, What/Why format) but delegates the actual executable workflow to references/workflow-steps.md. The inline content is more descriptive than instructive — the workflow summary is a one-liner pointing elsewhere rather than providing copy-paste-ready commands for each step. | 2 / 3 |
Workflow Clarity | The workflow is summarized as 'analyse → group → propose → name → stack → describe → remind' with a pointer to the full steps file, but the inline content lacks explicit validation checkpoints. There's no verify/fix loop after cherry-picking (only an anti-pattern saying 'don't leave conflicts unresolved'). The approval gate before branching is well-stated, but the actual sequenced steps with validation are deferred to a reference file. | 2 / 3 |
Progressive Disclosure | Excellent structure with a clear overview in the main file and well-signaled one-level-deep references (workflow-steps.md, anti-patterns.md, worked-example.md). Each reference has a clear description of what it contains and when to load it. The main file provides enough context to understand the skill without loading references. | 3 / 3 |
Total | 9 / 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.
Reviewed
Table of Contents