wb
Production-grade engineering skills for AI coding agents, built on Amazon Way of building services
Install / Use
npx skills add robisson/build-like-amazon-agent-skillsInstalls into whichever agent you are using.
Claude Commands
Claude Code slash commands
Quality Score
Category
Development & EngineeringSupported Platforms
Our assessment of wb
wb scores 64/100 on our quality scale, 3974th of 4,606 Development & Engineering skills we index.
Its Claude Commands is 3.3 KB long, split into 5 sections and no code examples: a solid amount of guidance for an agent.
It has no GitHub stars yet, so there is no community track record; judge it on its content.
Maintenance, license and trust
- The repository was last updated 21 days ago, so wb is actively maintained.
- No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
- Its trust signals score 80/100, with 2 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands.
Automated pattern scan on 2026-10-26. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
wb compared with similar skills
All 4 of these similar skills score higher than wb; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| wb (this skill)by robisson | 64 | 0 | 21d ago | Claude Commands |
| claude-memby thedotmack | 100 | 103.4k | today | CLAUDE.md |
| Agent-Reachby Panniantong | 100 | 103.3k | 1d ago | CLAUDE.md |
| taste-skillby Leonxlnx | 100 | 102.9k | today | CLAUDE.md |
| Understand-Anythingby Egonex-AI | 100 | 88.4k | today | CLAUDE.md |
Frequently asked questions
- How do I install wb?
- Run
npx skills add robisson/build-like-amazon-agent-skills. The install tabs above show the steps for each supported agent. - Which AI agents does wb work with?
- It is written for Claude Code, as a Claude Commands file. Other agents that read the same format can often use it too.
- Is wb safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. It declares no license and scores 80/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is wb still maintained?
- The repository was last updated 21 days ago, so wb is actively maintained.
Skill content
View source on GitHubWorking Backwards — Start Here
Path resolution: All
skills/,agents/, andpatterns/paths in this command are relative to the plugin root directory. Installed as a Claude Code plugin, that root is theBLA plugin root:line the plugin printed at session start; in a local install it is the working directory.
IMPORTANT: Each WB stage has a mandatory human approval gate. After completing each stage, present your work and ask the user to review before proceeding. Never skip stages or rush to PR/FAQ. The PR/FAQ is Stage 4 — it requires Stages 1-3 to be completed and individually approved first.
You are activating the Working Backwards meta-skill. This is Amazon's product discovery process: start from the customer and work backwards to the solution.
What to do
- Read the skill file at
skills/working-backwards/for full context. - Ask the user: "What product or feature are you thinking about?"
- Based on their answer, guide them through the five WB phases in order:
- Listen → Who is the customer? What do they need?
- Define → What is the specific problem we're solving?
- Invent → What is our proposed solution?
- Refine → Write the PR/FAQ document.
- Test → How will we validate and measure success?
- After each phase, STOP. Present the full output to the user. Ask them to review and approve before moving to the next phase. Do NOT proceed without explicit user confirmation.
- For each phase, load the corresponding skill from
skills/wb-*. - Produce artifacts in
.bla/working-backwards/— customer profiles, problem statements, solution sketches, and the final PR/FAQ.
Approval Gates (Non-Negotiable)
After completing each stage, present your output and ask: "Here's what I have for [Stage]. Does this look right? Anything to adjust before we move to [Next Stage]?" Wait for explicit approval. If the user asks for changes, revise and re-present. Only proceed after they confirm.
Key Principles
- Customer obsession: every decision traces back to a customer need.
- Write the press release BEFORE writing code.
- If you can't write a clear FAQ answer, the idea isn't ready.
- Iterate: it's normal to loop between phases multiple times.
Output
Create .bla/working-backwards/<feature-name>/ with:
customer-profile.mdproblem-statement.mdsolution-sketch.mdprfaq.mdsuccess-metrics.md
Flow metrics. /wb owns four of the six events for the wb phase — phase_started,
phase_completed, gate_approved and gate_rework — and appends each as one JSON line to
.bla/metrics.jsonl: phase_started once the change is classified as Medium or above,
phase_completed once the artifacts above are saved, and one gate_approved or gate_rework per stage
gate depending on whether the user approved the stage or sent it back. It emits spec_completed and
review_blocking_finding never — those belong to other commands, and no command emits an event another
one owns (owner table in docs/flow-metrics.md). Emit only at Medium and above; at Trivial and Small
emit nothing. If the line cannot be written — no writable tree, no docs/ directory, the adopter
declined — state in one line that the flow measurement for this phase was not recorded, and continue:
measurement never blocks a stage or a gate.
Related Skills
claude-mem
103.4kPersistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Agent-Reach
103.3kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
taste-skill
102.9kAgent skills that give coding agents taste: art direction, reference images and a browser check instead of generic AI pages
Understand-Anything
88.4kGraphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
