learn-from-pr

Analyzes a completed PR to extract lessons learned from agent behavior. Use after any PR with agent involvement - whether the agent failed, succeeded slowly, or succeeded quickly. Identifies patterns to reinforce or fix, and generates actionable recommendations for instruction files, skills, and documentation.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

85%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 highly actionable and workflow-sound: concrete gh commands, decision tables, a completion checklist, and worked examples for all three scenarios. Its weaknesses are mild redundancy and a monolithic structure that inlines example and mapping content that would be better split into reference files.

Suggestions

Move the three worked Examples (and optionally the pattern-to-improvement mapping tables) into a references/ file (e.g. references/examples.md) and link to it from a short Examples section, keeping SKILL.md as a lean overview.

Remove the duplicated learn-from-pr skill-vs-agent distinction: keep it once (in Integration) and drop the reference in Constraints and the final standalone paragraph.

Tighten Step 6 so it does not restate the Outputs section verbatim; a one-line instruction to present category, priority, location, change, and why suffices.

Dimension	Reasoning	Score
Conciseness	The body is dense and useful (pattern tables, commands, exact paths) with no explanations of concepts Claude already knows, but the learn-from-pr skill-vs-agent distinction is stated three times (Constraints, Integration) and Step 6 restates the Outputs structure, so minor trimming remains.	4 / 5
Actionability	It provides executable gh commands, a five-field recommendation template with exact file paths, and three fully worked example recommendations (one per scenario), which is copy-paste-ready guidance covering the common cases.	5 / 5
Workflow Clarity	Six clearly sequenced steps are backed by a Completion Criteria checklist and an Error Handling table with recovery paths ("PR not found \| Ask user to verify PR number"); the operation is read-only analysis, so the destructive-operation cap does not apply.	5 / 5
Progressive Disclosure	Sections are well-organized with clear headers, but the ~275-line body inlines material that belongs in separate reference files (the three worked Examples and the pattern-to-improvement mapping tables), with no bundle files or pointers, matching the "content that should be separate is inline" anchor.	3 / 5
	Total	17 / 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.

The description is strong: third-person voice, concrete capabilities, and an explicit 'Use after...' trigger clause covering all three outcome scenarios. It is highly complete and distinctive, with only minor room to grow in natural keyword synonyms.

Dimension	Reasoning	Score
Specificity	"Analyzes a completed PR to extract lessons learned", "Identifies patterns to reinforce or fix", and "generates actionable recommendations for instruction files, skills, and documentation" list multiple concrete actions covering the full analyze-to-recommend workflow, matching the comprehensive-coverage anchor rather than the minor-gaps anchor at 4.	5 / 5
Completeness	It explicitly states what ("Analyzes a completed PR to extract lessons learned... generates actionable recommendations") and when ("Use after any PR with agent involvement - whether the agent failed, succeeded slowly, or succeeded quickly"), which matches the anchor for both what and when with concrete trigger phrases.	5 / 5
Trigger Term Quality	Natural terms like "PR", "agent involvement", "failed, succeeded slowly, or succeeded quickly", and "lessons learned" give good coverage, but common synonyms a user might say (e.g. "postmortem", "retrospective", "learn from PR") are absent, so it falls short of the synonym-complete anchor at 5.	4 / 5
Distinctiveness Conflict Risk	The niche is narrow and clearly framed as post-completion learning analysis of agent behavior, explicitly distinct from PR-review or fix skills, matching the clear-niche/minimal-conflict anchor at 5 rather than the minor-overlap anchor at 4.	5 / 5
	Total	19 / 20 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: dotnet/maui
Path: .github/skills/learn-from-pr/SKILL.md
Commit: 7a5a5d6

Reviewed: 7 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.