Bruno Mazurok de Castro
Community@Mozurok · Brazil
Fhorja is a task-lifecycle engineering command system for AI-assisted software delivery, spanning discovery, planning, implementation, review, runtime verification, and delivery.
Agent Skills by Bruno Mazurok de Castro
Showing 98 vetted skills indexed across 1 GitHub repositories.
Mozurokautonomous-readiness
Validates project definition completeness before booting unattended autonomous runs.
Mozurokreview-hard
Reviews task code changes for correctness, safety, and maintainability risks before closure or PR prep.
Mozurokproblem-framing
Runs a socratic intake to produce a five-field BRIEF.md before task creation.
Mozurokexternal-research-fleet
Orchestrates parallel research workers to synthesize multi-angle findings into one grounded EXTERNAL_RESEARCH.md.
Mozuroktask-workspace
Provisions a dedicated git worktree and branch for an active task.
Mozuroktargeted-questions
Asks minimal high-value factual questions and persists answers in task repository files.
Mozurokjourney-map
Document multi-screen user journeys into structured journey spec markdown files.
Mozurokworkflow-guide
Explains which workflow command and editor mode to use now and why.
Mozurokimplement-slice-complement
Execute bounded micro-delta fixes within an existing implementation slice and record evidence in task state files.
Mozurokdirection-adjust
Record mid-task course corrections as numbered decisions in DECISIONS.md and update TASK_STATE.md.
Mozurokbackend-system-design
Generate a 12-section backend system-design RFC persisted as BACKEND_SYSTEM_DESIGN.md.
Mozurokai-feature-eval-harness
Designs dataset-backed evaluation plans for model-backed product features with measurable pass thresholds.
Mozurokweb-runtime-verify
Serves a built frontend on an ephemeral port and gates runtime behavior with a verification battery.
Mozuroktask-init-fleet
Decomposes a complex brief into parallel sub-tasks initialized by concurrent worker agents.
Mozurokextract-foundations-from-screens
Extract canonical design foundation tokens from SCREEN_SPEC markdown files.
Mozurokapply-sweep-triage
Persist sweep triage decisions into REVIEW_PREFERENCES.md for future finding suppression.
Mozurokapi-contract-review
Review API endpoint contracts for naming, methods, errors, auth, and pagination before implementation.
Mozurokslice-closure
Evaluates whether an implemented slice meets its exit criteria and records closure status in task state files.
Mozurokdesign-bootstrap
Extract Figma design tokens and components to scaffold design system documentation.
Mozurokbranch-commit
Generates branch names and commit messages grounded in the actual git diff.
Mozurokunity-scene-plan
Plan Unity GameObject hierarchies, component responsibilities, and networked authority before writing C#.
Mozurokautonomous-run
Executes approved implementation plans slice by slice under governor limits and human merge gates.
Mozurokincident-triage
Classify observed technical failures and route them to hotfix, slice, investigation, or escalation paths.
Mozurokcolor-contrast-architect
Audits design token foreground/background pairs against WCAG 2.2 contrast thresholds across themes.
Frequently Asked Questions About Bruno Mazurok de Castro
FAQPage SchemaWhat tasks can I accomplish with Mozurok's Fhorja skills?▼
You can run the full software delivery lifecycle: frame problems, initialize tasks, analyze impact, plan slices, implement approved changes, review code and security, verify web/app/API/Godot runtimes, audit accessibility and migrations, package PRs, and author postmortems and release plans.
Who are these skills designed for?▼
Software engineers, tech leads, and engineering teams using AI-assisted editors (Cursor modes: Ask, Plan, Agent, Debug) who want disciplined, artifact-grounded task execution. Reliability engineers, design-system maintainers, and security reviewers also get dedicated auditor personas.
How do the skills run in practice?▼
Each skill is a Markdown command with YAML frontmatter declaring its editor mode, consumed/produced context layers, and allowed tools (Read, Write, Edit, Bash, Glob, Grep, WebFetch, WebSearch). Skills read and write task artifacts like TASK_STATE.md and DECISIONS.md inside a projects/<client>__<project>/ folder structure.
What are the prerequisites and dependencies?▼
Skills require an active task folder created by task-init (project-bootstrap for new projects). Many are git-gated or multi-repo aware. Optional integrations include Figma MCP for design specs, Supabase or Postgres access for database introspection, and vetted MCP servers for messaging or issue trackers.
Are the skills safe to adopt from third parties?▼
The registry itself ships skill-vet and mcp-server-vet, read-only inspectors that scan third-party skills and MCP servers for exfiltration, secret access, prompt injection, and hidden Unicode, returning an install/decline/sandbox verdict for human approval before anything is trusted.