rootbr

rootbr/rooted

Plugin Claude Code

Plugin marketplace listing 2 plugins: code-quality, evidence-based-authoring.

rootbr/rooted

Settings file Claude Code

Agent settings declaring 1 allowed tools.

rootbr/rooted

Instructions file

Instructions for rootbr/rooted, covering rooted, orientation, evidence-based rule and quality gate.

rootbr/rooted

Plugin Claude Code

Deep Java code review with parallel specialist agents and proactive code cleanup.

rootbr/rooted

Skill Claude CodeCodex

Structural and boundary-level design rules: dependency direction, layering, component cohesion/coupling, plugin architecture, use-case isolation, boundary patterns, mutable-state segregation, concurrency isolation at the architectural seam. Applies when deciding where code lives, whether to split a module, how to…

rootbr/rooted

Skill Claude CodeCodex

Cleans up recently modified code: simplify, reduce nesting, remove redundancy, rename for intent, apply named refactoring moves, flag code smells, tighten error handling, keep tests clean. Runs proactively during any code change — feature implementation, bug fix, or refactoring session. Trigger phrases: 'clean this…

rootbr/rooted

Skill Claude CodeCodex

Review changes and create atomic commits following Conventional Commits. Use when the user asks to commit, wants help writing commit messages, or has staged/modified changes ready to land. NOT for rebasing, cherry-picking, amending, squashing, or other history rewriting; NOT for pushing or managing remotes.

rootbr/rooted

Skill Claude CodeCodex

Java code review — branches, PRs, diffs. Trigger on "review this branch", "find bugs", "what's wrong with this PR", "audit changes", "check diff", "review these Java changes", or any request to review Java code against project invariants. Any Java git repo, zero config required (optional config.md tunes invariants and…

rootbr/rooted

Plugin Claude Code

Evidence-based authoring: web research with source verification and AI context files grounded in traceable sources.

rootbr/rooted

Agent

Frontmatter indexer for the auditing-ai-context skill. Dispatched once by the audit workflow to scan the rule-card directory and return a structured index — ruleid, appliestotarget, checkkind, severitydefault, and path per card. It parses YAML frontmatter and emits JSON; it does not edit files.

rootbr/rooted

Agent

Read-only auditor for the auditing-ai-context skill. Dispatched by the audit workflow — to check a context file, skill, agent prompt, or KB card against a single rule and return schema-validated JSON patches. It proposes edits; it never applies them.

rootbr/rooted

Skill Claude CodeCodex

Critically appraise a finished research report, recommendation, or analytical document and return a ranked list of what genuinely needs re-checking, with revised confidence in its main conclusions. Runs seven independent critique methods as parallel background agents — evidence grading, source integrity…

rootbr/rooted

Skill Claude CodeCodex

Write, audit, and optimize AI agent instructions and the prose an agent produces. Use whenever the user creates, edits, or reviews CLAUDE.md, SKILL.md, agent definitions, or any agent-instruction file; writes system prompts, agent personas, or LLM instructions; audits knowledge-base cards or corpora (YAML facets…

rootbr/rooted

Skill Claude CodeCodex

Produce a maximum-breadth map of questions across any topic — covering all dimensions, perspectives, and angles so nothing important remains unasked. Use when the user wants to explore a topic comprehensively: learning a new field, brainstorming an incident, planning a book or project, researching options. Triggers on…

rootbr/rooted

Skill Claude CodeCodex

An AI reading assistant/companion for the user's own reading, studying, and absorbing of a book or material — methodology- and algorithm-driven active reading and note-taking. Two jobs: (1) set up a per-book workspace — classify the book, recommend how to read it, scaffold a customized CLAUDE.md; (2) run as companion…

rootbr/rooted

Skill Claude CodeCodex

Research any topic using web search with source verification and critical analysis. Handles narrow factual lookups directly and broad open-ended exploration by first invoking /discovering-subtopics to build a question map, then researching each question, synthesising, and appraising the report via /appraising-research…