Developer Tools●●●Banger
Ossature – Spec-driven code generation with LLMs
Fix task 14 of 30 without restarting—Cursor's all-or-nothing approach can't do this.
Big BrainSolve My Problem
beshrkayali
402mo ago
Back to browse
Crewly Codes – AI crew that specs, builds, and QAs your features
by digitalegoai·Feb 25, 2026·2 points·0 comments
AI Analysis
●●●BangerShip ItSolve My ProblemSlick
Specs before code—AI PM → DevCrew → QA, runs inside Claude Code today.
Strengths
- •Core insight addresses real pain: AI coding fails because specs are vague, not because models are bad
- •Portable MCP skills (markdown files) = zero lock-in, works in any MCP-compatible agent
- •Sequential handoff + shared context avoids parallelization chaos; concrete acceptance criteria forces rigor
Weaknesses
- •Adoption depends entirely on MCP ecosystem maturity and Claude Code's evolution
- •No evidence of real users or saved-time metrics yet; pricing / scale model unclear
- •Agents already integrated (DevCrew/QA) are black boxes—unclear how strict validation actually is
Category
Target Audience
Solo developers, small teams using Claude Code or Cursor
Similar To
Cursor · Continue.dev · GitHub Copilot Workspace
Post Description
Hey HN,
Crewly Codes is an AI product dev workflow where the AI specs the work before writing code.
Morgan (AI PM) turns your idea into structured specs — user stories, acceptance criteria, edge cases. You lock the spec, DevCrew builds it, QA validates it. Watch it happen or come back to finished code.
The thesis: most AI coding failures are spec failures. The AI builds the wrong thing because no one defined "done." Morgan forces that upfront.
Multi-agent but orchestrated — sequential handoffs with shared context, not parallel chaos.
Runs inside Claude Code and Codex. Free tier available.
Similar Projects
Developer Tools●●Solid
Spec-Driven Development Workflow for Claude Code
Context clearing between steps cuts costs, but hallucinated 2026 citations hurt trust.
Ship ItNiche Gem
sermakarevich
201214d ago
Developer Tools●●Solid
Generate Claude Code Workflows using Spec Driven Development approach
Spec-driven workflow generator for Claude Code when prompt-chaining dominates AI coding.
Big BrainNiche Gem
sermakarevich
507d ago
AI/ML●●Solid
Coding agent where a second agent QAs every PR in a real browser
QA agent drives real browsers via Browserbase to verify PRs before you wake up.
Bold BetShip It
kavin_key
1016d ago
Developer Tools●Mid
OctopusGarden – An autonomous software factory (specs in, code out)
Orchestrates AI agents to iterate code until tests pass—but StrongDM already ships this.
Bold BetShip It
foundatron
853mo ago
AI/ML●Mid
Claude Code skills for building LLM evals
Structured eval workflow for Claude Code when LangSmith and Braintrust already exist.
Niche GemShip It
paulaq
201mo ago