dld

Upgrade DLD framework files from latest template on GitHub.

Autonomous spec execution with subagents (planner, coder, tester, reviewer).

Autonomous spec execution with subagents (planner, coder, tester, reviewer).

Feature specification and research agent. Multi-agent with 4 scouts. Creates specs in ai/features/.

Feature specification and research agent. Multi-agent with 4 scouts. Creates specs in ai/features/.

Manual QA tester — tests product behavior like a real user, not code. Triggers on keywords: test, QA, check behavior, verify feature, manual testing, протестируй, потыкай, проверь как работает

Systematic code analysis (READ-ONLY). Light mode for focused checks, Deep mode for full forensics with 6 parallel personas.

Deep multi-agent bug analysis. 6 personas × N zones → findings report + inbox items for Spark. Triggers on keywords: bug hunt, deep analysis, охота на баги, баг-хант, командный аудит багов

Analyze diary entries → propose improvements to rules, agents, skills.

Seed ai/lessons/ from archive. Reads every BUG spec, understands the bug, writes structured lessons. Triggers on: seed lessons, засей уроки, заполни банк уроков, populate lessons bank.

Multi-agent review with 5 expert perspectives for architectural decisions.

Multi-agent review with 5 expert perspectives for architectural decisions.

Day-0 skill: extract idea from founder's head → structured files in ai/idea/.

Manual QA tester — tests product behavior like a real user, not code. Triggers on keywords: test, QA, check behavior, verify feature, manual testing, протестируй, потыкай, проверь как работает

System Architecture Board — multi-agent with 7 personas, cross-critique, LLM-ready gate.

Generate complete brand identity system for app launch. Strategy to store assets to marketing kit.

System Architecture Board — multi-agent with 7 personas, cross-critique, LLM-ready gate.

Deep multi-agent bug analysis. 6 personas × N zones → findings report + inbox items for Spark. Triggers on keywords: bug hunt, deep analysis, охота на баги, баг-хант, командный аудит багов

Analyze diary entries → propose improvements to rules, agents, skills.

Board of Directors — business architecture before system design. Multi-agent with cross-critique.

Write/modify code for autopilot tasks

Professional Excalidraw diagram generation from natural language or code analysis.

Detailed implementation planning with UltraThink - creates bite-sized tasks with full code examples

Automated documentation updates — CHANGELOG, README, docs. Fully automatic, no confirmations.

Brownfield lifecycle — reassess existing projects through Deep Audit, Architect recovery, Board re-prioritization, and Stabilization.

Code Quality Reviewer (Stage 2) - prevents tech debt and duplication

Isolated research agent for external sources (Exa + Context7).

Create new skills, modify existing skills, and measure skill performance. Triggers on keywords: create skill, new skill, write skill, eval skill, benchmark skill, optimize skill, improve skill, update CLAUDE.md, update rules, skill-creator

Run tests with Smart Testing and Scope Protection

Systematic code analysis (READ-ONLY). Light mode for focused checks, Deep mode for full forensics with 6 parallel personas.

Detailed implementation planning with UltraThink - creates bite-sized tasks with full code examples