Claude Skills by cryndoc

Author, run and repair human-written acceptance checks (образ результата) stored in acceptance/ACCEPTANCE.md — best-effort, prompt-held, no guarantees. Use when PM mentions "acceptance checks", "приёмка", "образ результата", "запусти приёмку", "acceptance run", "прогони приёмку", "почини по красным", or any request to record what "готово" means and hold the work against it. Trigger liberally — without it "готово" stays a matter of opinion; over-triggering is recoverable (PM can delete the file).

Capture a simple task that does not need planning (CHORE + TASK in readyToWork by default, or registration-only with --no-task). Use when PM mentions "simple task", "chore", "small task", "housekeeping", "quick task", "мелкая задача", "чора", or any request to log routine work without PRD/SPEC overhead. Trigger liberally — under-triggering pushes tiny work into freeform chat; over-triggering is recoverable (PM can delete).

Autonomous work — find and execute ready tasks

Record a technical-debt item (DEBT-NNN) — registration only by default, or register + auto-generate a fix TASK with --task. Use when PM mentions "add tech debt", "record tech debt", "technical debt", "tech debt item", "refactor tracking", "технический долг", "запиши техдолг", or any request to capture deferred refactoring / cleanup work. Trigger liberally — under-triggering loses debt visibility; over-triggering is recoverable (PM can delete or defer).

Register a defect (BUG-NNN) and auto-generate the fix TASK so the bug enters the normal implement/review flow. Use when PM mentions "file a bug", "report defect", "bug report", "register defect", "report a bug", "заведи баг", "баг-репорт", or any request to capture a defect for tracking. Trigger liberally — under-triggering leaves bugs in chat where they get lost; over-triggering is recoverable (PM can delete the BUG artefact).

EXPERIMENTAL (Claude Code only). Apply a SPEC increment to a single LIVING architecture corpus under docs/architecture/ instead of a per-SPEC silo package — treats docs as event-sourcing (SPEC = commit, corpus = working tree), so C4 Context/Container, glossary and the data model stay system-wide and never drift across SPECs. Use when PM mentions "living corpus", "single architecture", "corpus mode", "merge design into the corpus", "one architecture for all SPECs". Opt-in behind settings.exper...

Create a doc-as-code design package from a PRD or SPEC. Conditionally generates C4 diagrams (Context/Container/Component), sequence diagrams, ER diagram + Data Dictionary, OpenAPI 3.0, AsyncAPI 3.0, ADRs, domain glossary, state diagrams, and deployment view as Mermaid-rendered Markdown files. Use when PM mentions "design", "architecture diagrams", "doc-as-code artifacts", "C4", "ERD", "OpenAPI spec", "AsyncAPI", "event-driven", "Kafka", "message broker", "sequence diagram", "state machine", "...

Diagnose Polisade Orchestrator project health

Create a lightweight Feature Brief (FEAT-NNN) for incremental product work that does not warrant a full PRD. Use when PM mentions "create a feature", "add feature brief", "new feature work", "feature request", "add FEAT", "заведи фичу", "новая фича", or any request to capture a small-to-medium feature without PRD-level process. Trigger liberally — under-triggering forces the agent to improvise freeform docs; over-triggering is recoverable (PM can delete or upgrade to PRD).

Implement TASK

Verify that /polisade:init produced a canonical project structure (valid .state JSON, real polisadeVersion/schemaVersion, context file present) rather than content reconstructed from memory. Use when PM mentions "verify init", "check init", "validate project structure", "проверь init", "проверь инициализацию", or right after /polisade:init to confirm the generated files are byte-correct before reporting success.

Scaffold the full Polisade Orchestrator framework (.state, docs, templates, backlog, CLAUDE.md) into the current project. Use when PM mentions "init project", "initialize", "scaffold project", "set up Polisade", "bootstrap project", "инициализируй проект", "разверни структуру", or any request to bring up the autonomous development structure in a fresh or existing repository.

Lint and validate Polisade Orchestrator skill definitions

Upgrade PROJECT_STATE.json schema to current version

Provider-agnostic PR operations (GitHub / Bitbucket Server) for PM — create, list, view, diff, merge, comment, close, whoami. Use when PM mentions "open a PR", "create pull request", "push and open PR", "submit for review", "commit and open PR", "merge PR", "review PR status", "закоммить и сделай pr", "сделай пиар", "сделай pr", or any request to operate on GitHub / Bitbucket Server pull requests. Trigger liberally — skipping forces the agent to improvise with ad-hoc REST calls that bypass OP...

Create a full Product Requirements Document (PRD-NNN) via a structured PM interview — for large initiatives that justify the overhead. Use when PM mentions "create PRD", "write product requirements", "PRD document", "product brief", "requirements doc", "напиши PRD", "создай PRD", or any request to capture a full product vision before handing to design / engineering. Trigger liberally — under-triggering forces the agent to write freeform docs that drift from the canonical PRD template; over-tr...

Show and manage open questions across all artifacts

Best-effort reconciliation of the living architecture corpus against the code: the model reads docs/architecture/ plus the code and lists divergences with coordinates. Advisory only — an opinion of the model, never a verified verdict, no gates and no stamps. Nothing is applied automatically: corpus edits are a separate confirmed step written through the corpus primitive.

Run a quality review on an open pull request (by PR number or linked TASK), externally via codex or internally via a clean-context self-review, and post the verdict. Use when PM mentions "review PR", "PR review", "review pull request", "quality review", "review this pr", "сделай ревью pr", or any request to evaluate an open PR before merge. Trigger liberally — under-triggering lets PRs land without a second look; over-triggering is recoverable (PM can ignore the review comments).

Run a second-opinion quality review on a completed TASK — externally via codex or internally via a clean-context self-review — and gate the merge decision. Use when PM mentions "review TASK", "second opinion", "check my implementation", "codex review", "self-review", "review task", "сделай ревью", or any request to get another pass over TASK work before marking it done. Trigger liberally — under-triggering lets half-baked TASKs land without a second look; over-triggering is recoverable (PM ca...

Build an implementation PLAN (PLAN-NNN) from a technical SPEC — milestones, phases, dependencies — via a clean-context subagent. Use when PM mentions "create PLAN", "plan from SPEC", "roadmap", "implementation plan", "milestones", "create roadmap", "сделай план", or any request to sequence SPEC work into execution phases. Trigger liberally — under-triggering lets the agent jump straight to TASKs without a milestone view; over-triggering is recoverable (PM can delete or regenerate).

Generate a technical SPEC (SPEC-NNN) from an existing PRD or Feature Brief via a clean-context subagent. Use when PM mentions "create SPEC", "write spec", "turn PRD into SPEC", "functional spec", "specification from PRD", "создай spec", "напиши спеку", or any request to translate product intent into engineering-ready specification. Trigger liberally — under-triggering forces the agent to improvise design/API calls in chat without the canonical SPEC template; over-triggering is recoverable (PM...

Create a timeboxed research SPIKE-NNN for answering an open technical / product question before committing to implementation. Use when PM mentions "research task", "spike", "investigation", "research question", "explore a problem", "ресёрч", "спайк", or any request to investigate an unknown before locking a design. Trigger liberally — under-triggering forces the agent to speculate without recording the research trail; over-triggering is recoverable (PM can close the spike early).

Show project status

Rebuild PROJECT_STATE derived fields from artifact files

Decompose a PLAN / SPEC / FEAT / BUG / DEBT / CHORE into atomic TASK-NNN items ready for the implement flow, via a clean-context subagent. Use when PM mentions "create tasks", "generate TASKs", "break down into tasks", "tasks from PLAN", "break this into tasks", "декомпозиция", "сделай таски", or any request to explode upstream artefacts into executable work. Trigger liberally — under-triggering forces ad-hoc task-creation in chat that drifts from backlog conventions; over-triggering is recov...

Answer PM questions and unblock tasks