Intent-driven AI-First development methodology for Claude Code and Codex — Project Knowledge, user-spec planning, focused execution, and evidence-gated reviews.
This repository also configures its own agents. See what molyanov-ai-dev tells them →
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Researches codebase for a feature: files, patterns, tests, integrations, risks. Creates or deepens code-research.md for user-spec-planning.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews code quality after implementation, including localized edits, features, refactors, cross-file changes, generated artifacts, and behavior changes.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews Project Knowledge for demonstrated content gaps, generic tutorial material, duplication, stale facts, misplaced content, and contradictions.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews changed or existing project infrastructure, CI/CD, deployments, release artifacts, recovery, retention, and monitoring for demonstrated failures. Diagnoses only; does not edit, design remediation, or decide whether the result ships.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews user-spec interview evidence against project knowledge and code research to diagnose unresolved requirements before drafting. Use when: interview cycles are complete and completeness must be checked before drafting. Diagnoses gaps only; follow-up questions, requirement choices, and drafting decisions are out…
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews visual fidelity and responsive layout after layout-writing work. Uses supplied source and captured evidence without redesigning the interface or modifying code.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews LLM prompts for demonstrated clarity, framing, structure, compression, context, and prompt-injection risks against prompt-master principles.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Audits changed code and relevant callers for demonstrated OWASP Top 10 vulnerabilities, exposed secrets, and applicable dependency risks.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Verifies factual user-spec claims against the current codebase, including paths, symbols, dependencies, integrations, behavior, and project patterns. Use when: validating the factual-codebase lane; solution adequacy and document quality are out of scope.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews a skill's form against skill-master: frontmatter, routing, package structure, references, line limits, instruction style, and applicable skill-type conventions.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews a skill for executable logic on required paths: contradictions, dead ends, missing required results or state, and ordering failures.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews a skill for demonstrated over-engineering in rules, exceptions, scripts, phases, thresholds, options, and references.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews written test code before or after implementation and diagnoses demonstrated gaps in scenario coverage, assertions, boundaries, and test quality.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews user-spec feasibility, proportionality, architecture compatibility, over- or under-engineering, and concrete simpler alternatives against the current project. Use when: the user-spec is ready for pre-approval solution review; document quality and factual codebase claims are out of scope.
pavel-molyanov/molyanov-ai-dev
Agent Claude Code
Reviews user-spec document quality: structure, interview coverage, acceptance-criteria testability, edge-case presence, contradictions, and template compliance. Use when: the user-spec is ready for pre-approval document review; solution adequacy and factual codebase claims are out of scope.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: