Skip to content

ilang-ai/autocode / skills/product-study

product-study

Unverified

Study an existing product from its public pages and turn it into an evidence-backed blueprint plus an MVP cut that AutoCode can build and ship. Use this whenever the user wants something "like X", "similar to X", "my own version of X", "a clone of X", names a competitor they admire, or pastes a product URL and asks to build it — even when they never say "research" or "analyze". Public sources only (official site, docs, pricing, changelog, public repos, app-store listings, public reviews); no logins, captures, decompiling, or copying of code, copy, images or brand.

references/boundaries.md

View on GitHub
# Boundaries: what product-study does not do, and what to say instead

These are product rules, not a lecture. Each one exists because crossing it
would make the user's product unsafe to ship or would stop being "their own".
When a request lands on one of them, answer in one sentence, offer the in-scope
alternative, and keep working. Never stop the task over a boundary, and never
explain the rule at length unless asked.

## 1. No reverse engineering

Decompiling, unpacking, de-obfuscating, reading minified bundles or source maps
to recover logic, intercepting traffic, probing APIs, instrumenting the app.

Why: it is the part of "learn from a product" that carries legal exposure (terms
of service, anti-circumvention, trade secrets) and it produces implementation
details, not structure. A product built from someone's recovered internals is
not the user's own.

Say: *"I work from the public pages; here's what they tell us about how it's built."*

## 2. No account areas

Logging in — with the user's credentials or any other — to walk the original's
interface, export data, or read private docs.

Why: account areas are governed by terms the user accepted personally; pages
there are not public and often contain other people's data.

Say: *"I don't use account areas. If you describe what you see in yours, I'll
use that as a medium-confidence source."* A screenshot the user takes of their
own account is fine to read for structure.

## 3. No copying of the skin

Interface layouts, visual design, colors, icons, typography, copy, screenshots,
images, logo, product and feature names that are trademarks, slogans.

Why: copyright and trade dress live here, and it is also the part that would
make the user's product look like a knock-off. The structure is the lesson; the
skin is theirs to design.

Say: *"We take the structure and design our own look; section 10 lists what we
won't copy."* If the user insists on a look-alike, build a clean original design
that serves the same flows and say in one line why it is not a copy.

## 4. No code transplant

Copying code from the original's repositories or bundles into the build.

Why: licenses (copyleft in particular) and provenance. A product with one clear
origin is easier to sell, open-source, or audit later.

Say: *"I'll write this part fresh; the repo told us the shape, the code is ours."*

## 5. No non-public material

Leaked source, internal documents, beta or partner material, anything the user
"found" rather than something that was published.

Why: using it would taint the whole blueprint — every other source is clean, and
one bad one makes the document indefensible.

Say: *"I'll leave that out and work from the public pages."* Do not read it even
if it is pasted into the conversation; do not quote from it.

## 6. No bulk or evasive collection

Bypassing robots.txt, rate limits or anti-bot checks; scraping user-generated
content in bulk; fetching with a disguised user agent.

Why: the tool identifies itself and asks permission the way a browser-driven
person would; that is what keeps the "public page" claim true.

Say: *"That path is closed to crawlers; I'll use the pages that are open."*

## In-between cases

- **The product has no public docs at all.** Build a thin blueprint from the
  homepage, pricing, store listing and reviews; most lines will be `{medium}`
  or `{guess}`; say so in section 0. The MVP cut will lean on the user's goal.
- **The user is the owner of the original** (rebuilding their own product).
  Their own repositories, docs and accounts are in scope — cite them as such.
  The skin rules still apply if the new product is meant to be a separate brand.
- **The original is open source under a permissive license and the user wants
  to fork it.** That is a different task (fork and modify), not a study; say so
  and ask which they want.
- **The user asks for a feature comparison between two products.** In scope;
  study both, one slug each, then compare in the chat or in a third document.
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 198kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 198kUpdated Claude CodeCodexVerifiedInstall