Plugin Ecosystem

pensive

Reflective code review toolkit with structured analysis, domain expertise, and intelligent review orchestration.

New

Source: claude-night-market marketplace Version: 1.4.3 Author: Alex Thola

Key Commands

Command	Description	When to Use
`/full-review`	Comprehensive multi-dimensional code review with intelligent skill selection	When you want a thorough review covering all angles
`/architecture-review`	Principal-level architecture assessment against ADRs and design patterns	Before major refactors or design decisions
`/api-review`	API design and contract review	When designing or changing APIs
`/bug-review`	Systematic bug detection with language-specific expertise	When hunting for bugs in a codebase
`/test-review`	Test quality and coverage review	When evaluating test suite quality
`/rust-review`	Rust-specific code review (safety, performance, idioms)	When reviewing Rust code
`/math-review`	Mathematical correctness review	When code involves math or algorithms
`/makefile-review`	Makefile analysis and best practices review	When reviewing build configurations
`/shell-review`	Shell script review (safety, portability, best practices)	When reviewing shell scripts
`/skill-review`	Claude Code skill quality review	When reviewing plugin skills
`/skill-history`	View skill evolution and change history	When understanding skill changes over time
`/refine-code`	Iterative code refinement across 6 quality dimensions	When improving existing code quality

Skill	When It Activates
api-review	During API design review
architecture-review	During architecture assessment
bug-review	During bug detection analysis
code-refinement	During iterative code improvement
fpf-review	During functional perspective framework review
makefile-review	During Makefile analysis
math-review	During mathematical correctness review
rust-review	During Rust code review
safety-critical-patterns	When safety-critical code patterns apply
shell-review	During shell script review
test-review	During test quality assessment
unified-review	During comprehensive multi-dimensional review

Agent	Purpose
code-reviewer	Multi-file code review producing structured findings
architecture-reviewer	System design and ADR compliance assessment
rust-auditor	Rust-specific security and performance audit
code-refiner	Iterative code refinement across quality dimensions

The most common pensive workflow for reviewing changes before merging:

Start review — Run /full-review on your changes. Pensive analyzes the codebase to determine which review skills are most relevant (e.g., if there are shell scripts, the shell-review skill activates).
Review findings — The unified-review skill produces structured findings grouped by severity and category. Each finding includes the file, line, issue description, and suggested fix.
Refine — For any issues found, run /refine-code to iteratively improve the code. The code-refiner agent works across 6 quality dimensions.

When preparing for a major refactor or evaluating system design:

Review architecture — Run /architecture-review. The architecture-reviewer agent performs a principal-level assessment, checking against ADRs (Architecture Decision Records) and established design patterns.
Check compliance — The review covers layering violations, dependency direction, API contract consistency, and pattern adherence.
Act on findings — Use the findings to guide your refactoring plan, updating ADRs where needed.

/full-review is the best starting point — it automatically selects the right specialized skills based on your codebase.
Domain-specific reviews (/rust-review, /shell-review, /math-review) go deeper than /full-review in their specific area.
The fpf-review skill (Functional Perspective Framework) provides a unique angle that examines code from the perspective of each function's purpose.
/refine-code works iteratively — run it multiple times to progressively improve code quality.