matreshka-agent

Plugin matreshka-agent from marketplace matreshka-agent in GeneralAi360/matreshka-agent. Commands read from the repo's manifest — 21 days ago.

Skills (11)

• building-end-to-end

  Build or extend a substantial app, site, bot, integration, or feature from a plain-language request by selecting a simple interaction mode and project scenario, then entering Matreshka Agent's end-to-end controller. For user-facing products, automatically include Design Intelligence so UX/UI directi

• debugging-systematically

  Diagnose the root cause of a software defect, failing or flaky test, unexpected runtime behavior, performance regression, or integration error using controlled evidence. Use when the cause is unknown or disputed, before proposing or implementing a fix. A diagnosis request does not authorize code cha

• designing-product-experience

  Design or reconcile the UX/UI direction for a website, application, mobile interface, dashboard, onboarding flow, or other user-facing product. Use when the user wants a beautiful coherent interface, does not know what visual direction they want, needs a durable root DESIGN.md, wants multiple intera

• finishing-development-work

  Safely finish verified development work by preserving state, resolving required design and documentation drift, selecting and executing only authorized Git or handoff actions, and producing an exact continuation record. Use after review and fresh verification when the user wants to keep local work,

• implementing-with-tests

  Implement a bounded software feature or bug fix through a focused test-first cycle and produce evidence for review. Use when an approved task requires code/configuration changes, regression evidence, focused RED/GREEN proof, and nearby regression checks. The task may carry Project Intelligence (`ARE

• orchestrating-subagent-work

  Orchestrate software-development execution with repository inspection, source-intent traceability, Project Intelligence, Design Intelligence for UI-bearing work, security-by-design specification, planning, bounded subagents, review, verification, recovery, and handoff. Use directly when coordinating

• planning-software-work

  Convert a confirmed software specification or bounded change into an implementation plan with requirement coverage, project-area/interface routing, design-contract routing when UI is affected, security controls, small ordered tasks, exact repository paths, focused RED/GREEN checks, review gates, and

• reviewing-agent-work

  Independently review an agent-produced scoped diff and its evidence for specification compliance, correctness, frozen cross-area interface compatibility, design-contract consistency when UI changes, security, isolation, leakage, maintainability, and test sufficiency. Use after implementation or a re

• specifying-software-work

  Turn a raw product idea, new feature, ambiguous requirement, risky behavior change, or several architectural/experience options into a confirmed, security-by-design software specification before implementation. When a controller supplies an accepted Design Intelligence outcome for UI-bearing work, p

• verifying-development-work

  Verify development claims against current repository state and acceptance criteria using fresh, proportionate command and interaction evidence. Use after implementation or fixes, before declaring a task, phase, branch, or handoff complete, or when asked to prove that existing work functions. For UI-

• writing-portable-agent-prompt

  Write or improve a self-contained prompt for a software-development coding agent, especially when the prompt is drafted in one AI and executed in Codex, Claude Code, Cursor, or Antigravity. Use when the user wants the prompt prepared but not executed, including read-only investigation, implementatio

Similar plugins