.Agent

Guidelines for fast-tracking trivial tasks (typos, UI tweaks, simple bugfixes).

Use when high-assurance reliability is required (Bug fixes, Critical Features, Quality Hardening).

Use when a run of a workflow, skill, command, or test produced errors or friction worth keeping, or when executing the end-of-run Retro Global Protocol — collect findings into the feedback inbox, triage them (defect / work-item / noise), and file defects into the known-issues ledger or work-items into the backlog. NOT for fixing already-filed issues (/heal-issues). Triggers: "собери фидбек по прогону", "file run errors", "retro this run", "/run-feedback".

Use when brainstorming ideas, exploring solutions, generating options, clarifying requirements, designing architecture, or answering open-ended "how should we" and "what are the options" questions — even without the word "brainstorm".

Use when the user wants to customize Gemini CLI behavior using hooks (events, blockers, loggers).

Use when creating new Agent Skills, upgrading existing skills, running evals to test a skill, benchmarking skill performance, or optimizing a skill's description for better triggering accuracy. Guidelines for Gold Standard skill structures.

Use when performing VDD adversarial review with an opt-in sarcastic, provocative delivery style — a stylistic skin over vdd-adversarial mechanics (exhaustive reporting + objective bar).

Use when performing security vulnerability assessment (OWASP, secrets, dependencies, IaC, LLM, API, MCP/agentic) or when "thinking like a hacker" to find exploits.

Use when you need to audit, fix, or improve an existing agent skill to meet Gold Standard compliance.

Test-Driven Development with Stub-First approach.

Fundamental principles of agentic development: atomicity, traceability, stub-first, minimizing hallucinations.

Use when performing Verification-Driven Development with adversarial approach. Actively challenge assumptions and find weak spots.

Standards for Technical Specifications (TASK) and Use Cases.

Complete protocol for archiving TASK.md and PLAN.md (lockstep) with ID generation. Single source of truth for archiving.

Guidelines for designing scalable and modular system architecture.

Use BEFORE authoring a specification (TASK, ARCHITECTURE, PLAN, task files, issue records), a CHANGELOG entry or release notes, and AFTER to audit one that reads as an essay: long sentences, evaluative words, maxims, personification, metaphors, reasoning inside requirements, emoji as severity. Self-probing scanner; never changes the document's language. Triggers: «формализуй задачу», «проверь регистр», «напиши CHANGELOG», "formalize this spec", "reads like an essay", "write the spec", "release notes".

Rules for managing local .AGENTS.md and global artifacts (TASK.md, PLAN.md, ARCHITECTURE.md, KNOWN_ISSUES.md, BACKLOG.md).

Use when creating, formatting, or adding an entry to a project's thin-index ledger — `docs/KNOWN_ISSUES.md` + `docs/issues/` for defects, or `docs/BACKLOG.md` + `docs/backlog/` for work-items — or seeding either in a new project: shared index-over-records mechanics, per-registry frontmatter schema, prefix→category table, status/severity/effort vocab, index-line format, record-file recipe.

Skill loading tiers: TIER 0 (always), TIER 1 (phase-triggered), TIER 2 (extended). Defines when to load which skills.

Centralized list of commands safe for auto-execution without user approval. Single source of truth.

Persist and restore agent session context (Mode, Task, Summary) to survive resets.

Auto-update .AGENTS.md files based on code changes.

Guidelines for the Developer role: strict adherence, no unsolicited refactoring, documentation, security.

Validates conformance between TASK.md (RTM) and PLAN.md (Atomic Checklists).

Structured checklist for code review: bugs, style, performance, security, docs.

Detailed checklist for verifying Development Plans.

Use when gathering and refining requirements into a structured TASK.

Standards for Development Plans (PLAN.md) and Detailed Task Descriptions (TASK.md).

Guidelines for managing the Product-to-Technical handoff, including Quality Gates and BRD Compilation.

Detailed checklist for verifying Technical Specifications (TASK).

Decision tree for breaking down tasks and prioritizing work.

Guidelines for creating and refining Product Vision and Strategy.

Guidelines for implementing effective and maintainable tests.

Standards for code documentation, comments, and artifact updates.

Core structure for Architecture documents. For full templates with examples, load architecture-format-extended.

Regenerate architecture documentation from codebase analysis.

Use when performing OWASP security critique in adversarial style (optional sarcastic skin). Part of VDD Multi-Adversarial pipeline.

Stage Cycle pattern and dispatch table for Orchestrator compression. Defines common Init→Review→Revision flow.

Use when decomposing tasks into parallel sub-tasks or spawning sub-agents. Vendor-agnostic core; load a per-vendor reference for concrete tool names, directory conventions, and invocation syntax.

Detailed checklist for verifying System Architecture and Data Models.

Performance critic in adversarial style (optional sarcastic skin). Part of VDD Multi-Adversarial pipeline.

Guidelines for creating Solution Blueprints (UX Flows, ROI, Risk Registers) for BRD preparation.

Guidelines for conducting Market Research (TAM/SAM/SOM) and Competitive Analysis.

Meta-skill for verifying Framework Upgrades (Gemini Self-Improvement).

Guidelines for managing and prioritizing the Product Backlog using WSJF.