Skip to main content

pr-review-toolkit

Specialised review agents for comments, tests, error handling, type design, code quality and simplification, orchestrated by /review-pr.

Reference​

FieldValue
Namepr-review-toolkit
Version1.0.0
CategoryProductivity
LicenceNot declared
Sourcecatalog/navin/builtin_plugins/pr-review-toolkit/plugin.json
READMEcatalog/navin/builtin_plugins/pr-review-toolkit/README.md
Digestsha256:3dd407df62581242fc7c4506ddf3dbbea288291350295e55b0881f1cd0bf172e
plugin.jsonagents, category, commands, description, name, skills, steps, version

Commands​

NameArgumentsDescription
/review-pr[comments|tests|errors|types|code|simplify|all] [parallel]Review the current change or pull request with specialised agents (code, tests, errors, types, comments, simplify)

Agents​

NameToolsModelDescription
code-reviewerread_file, grep, find_files, list_dir, code_index, lsp, code_review, web_fetch, web_search-Reviews code for bugs, logic errors, security flaws, quality problems and project-convention breaches, scoring each finding for confidence and reporting only the high-priority ones. Use when code was written or changed and needs a review before it is merged.
comment-analyzerread_file, grep, find_files, list_dir, code_index, lsp, code_review-Checks code comments and docstrings for accuracy against the code, completeness and long-term value, and flags comment rot. Use when documentation comments were written or changed, before finalizing a pull request that touches comments, or when auditing existing comments for technical debt.
pr-test-analyzerread_file, grep, find_files, list_dir, code_index, lsp, code_review-Reviews the test coverage of a change for behavioural completeness, critical gaps and test quality, rating each missing test by how much it matters. Use when a pull request is opened or updated, when new behaviour was added, or before marking a change ready for review.
silent-failure-hunterread_file, grep, find_files, list_dir, code_index, lsp, code_review-Hunts silent failures, swallowed errors, overly broad catch blocks and unjustified fallbacks in changed code, and reports each with its severity, hidden errors and a corrected example. Use when a change adds or modifies error handling, try/except or catch blocks, fallback logic, retries, or anything that could suppress an error.
type-design-analyzerread_file, grep, find_files, list_dir, code_index, lsp, code_review-Analyses the design of new or changed types for invariants, encapsulation and enforcement, and rates encapsulation, invariant expression, usefulness and enforcement from 1 to 10. Use when a type, class, model or schema is introduced, when a pull request adds data types, or when an existing type is refactored.
code-simplifierread_file, grep, find_files, list_dir, code_index, lsp, code_review, apply_patch, edit_file, verify, test_run-Simplifies recently written or modified code for clarity, consistency and maintainability while preserving its exact behaviour, following the project's own conventions. Use when a coding task or a logical chunk of code is finished, after a review has passed, or when code works but feels more complex than it needs to be.

Skills​

NameDescription
code-reviewerReview code changes for bugs, security issues, regressions, and maintainability.
test-generatorGenerate and run unit, API, and E2E tests (including Playwright) that match project conventions.

Hooks, MCP servers and settings​

plugin.json declares no hook, no MCP server and no setting: the plugin is made of the components above only.

Workflow​

StepUse
review a change from every angle/review-pr
review local changes against the guidelinesagent:code-reviewer
check comments and docstringsagent:comment-analyzer
check test coverageagent:pr-test-analyzer
hunt silent failures in error handlingagent:silent-failure-hunter
review new types and invariantsagent:type-design-analyzer
simplify code after the reviewagent:code-simplifier
write the missing testsskill:test-generator

Enable it in a project​

Activating pr-review-toolkit creates 10 activations in the project: the plugin itself, 2 skills, 1 command and 6 agents. Removing the plugin activation removes them all.

curl -X POST <control-plane-url>/v1/projects/<project>/navin-library/activations \
-H "Authorization: Bearer <token>" -H "Content-Type: application/json" \
-d '{"kind": "plugin", "name": "pr-review-toolkit"}'

<project>: id or slug of the project. <token>: API key with the editor role on the project.

client.navin_library.activate("<project>", "plugin", "pr-review-toolkit")