ai-assisted-engineering: a documentation repository, not a tool
A Guide To AI-Assisted Software Engineering
At a glance
- What is it?
- John Crickett's ai-assisted-engineering is a Markdown index of how-to guides, tool lists and reading material for engineers adopting LLM coding assistants. The judgement: it is a pointer collection, and its value depends entirely on whether the linked pages stay current.
- Who is it for?
- Adopt it as a starting index if you are a working engineer or an engineering lead who wants a single README that points at AGENTS.md conventions, coding agent skills and prompt engineering material, and you are willing to open each linked file yourself. Do not adopt it if you want something you can install, version and pin, because the repository ships no code, no schema and no release artefacts.
- Can I use it commercially?
- Not without permission. GitHub finds no licence file in the repository, and without a licence all rights are reserved by default: you may read the code but not reuse it. Check the README, or ask the authors, before using it.
- Is it still maintained?
- Activity is slowing. The repository last received commits 6 months ago.
- What is it written in?
- GitHub does not report a main language for this repository.
Answers come from the project's GitHub data, last synced on September 15, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
What problem a guide repository solves for an AI coding team
The README opens by stating that AI tools for developing software based on LLMs are becoming prolific and that engineers and engineering leaders need to adapt. That framing is the whole product thesis. The repository does not generate code, does not wrap a model and does not run an agent. It is a Markdown tree whose job is to answer the question a team asks the week it decides to standardise on AI assistants: what do we write down, and in what order?
The audience is narrow and specific. It is the engineer who has already used a coding assistant and now needs to explain to colleagues how prompts, agent instruction files and code review tooling fit together. It is also the engineering manager who has to pick tools without reading forty vendor pages. The README addresses both by splitting the tree into three top-level groups: How To Guides, Guides, and Resources. That split is the useful part. It separates procedural material from conceptual material from link lists, which is more than most AI-tooling collections do.
How the repository is laid out: three folders and a link list
The README is an index. The Sections block links to HowTo/README.md, Guides/README.md and Resources/README.md, and then expands each one inline.
HowTo currently lists a single entry: requirements.md. Guides lists three: agents.md, skills.md and prompt-engineering.md. Resources is the widest branch, with subfolders for CodingTools, CodeReviewTools, VendorGuides, Courses, Blogs, Talks and Books.
Two structural details are worth noting. First, the README links Resources/Courses/README.md twice, once as AI Assisted Coding Courses and once as Useful AI Related Courses. That is either a duplication that should be merged or two distinct pages that need distinguishing labels. As published, a reader cannot tell which. Second, every heading in the tree is a link to a file that is not shown in the supplied material, so the depth of each page is unknown from the README alone. The repository layout suggests a documentation-only project: no package manifest, no build config, no source directory is referenced anywhere in the README.
The Guides branch is where the actual engineering content sits. A page on AGENTS.md implies conventions for the instruction files that coding agents read from a repository root. A page on coding agent skills implies the skill or capability units that agents load on demand. A page on prompt engineering implies text-level technique. Those three topics map to three different layers of an AI coding setup (repository-level instructions, per-task capability, and the prompt itself), and grouping them into one folder is a reasonable taxonomy.
Reading it: there is nothing to install
There is no install step, no CLI and no configuration. Getting value out of this repository means cloning or browsing it and opening the Markdown files:
git clone https://github.com/JohnCrickett/ai-assisted-engineering.git
Then read the entry points in the order the README presents them:
HowTo/README.md, then HowTo/requirements.md Guides/README.md, then Guides/agents.md, Guides/skills.md, Guides/prompt-engineering.md Resources/README.md, then the subfolders for coding tools, code review tools, vendor guides, courses, blogs, talks and books
If you want to use the Guides folder as internal onboarding material, the practical move is to copy the relevant Markdown into your own repository and edit it, because the upstream project has no templating, no variables and no config keys to adapt. There is no AGENTS.md in this repository itself as far as the README shows, which is a small irony given that the repository documents AGENTS.md as a topic. Anyone treating this as a reference implementation of its own advice will be disappointed.
One thing the README does not state anywhere in the supplied text: the licence. The repository metadata also does not record one. Before copying pages into a company wiki, that gap matters, and it is the first thing to check in the tree.
Where a link-list repository breaks down
The failure mode of this format is link rot and staleness, and the structure makes it worse than it needs to be. Every substantive claim in the repository lives one hop away in a file the README does not summarise. If a vendor guide URL dies, or a coding tool is renamed, the README still reads as complete because the index entries do not carry dates, version numbers or status markers.
The Resources branch is the most exposed. Tool directories, code review tool lists, course lists and blog lists all decay on a scale of months in this field. The repository's last push timestamp is 2026-03-13, which tells you when something changed but not which pages were touched. Nothing in the supplied material indicates a changelog or a review cadence.
A second limitation is scope. The HowTo branch has one page. If you came for a full workflow, from writing requirements through to reviewing AI-generated diffs, the repository does not currently provide it. The Guides folder covers instruction files, skills and prompting, which are three inputs to an AI coding workflow, not the workflow itself.
A third is the wrong-tool case. If your problem is that your agent produces inconsistent code across a monorepo, this repository will not fix it. It will tell you that AGENTS.md exists and point you at a page about it. The fix lives in your own repository, in the instruction file you write and the skills you define. This project is reading material for the person writing those files, not a component of the system that consumes them.
How it differs from an agent framework or a prompt library
The nearest alternative in kind is an agent framework or prompt library, for example a repository that ships runnable prompt templates, tool definitions and orchestration code you import into a project. The difference in approach is concrete: a framework gives you artefacts you execute and version, while ai-assisted-engineering gives you prose you read. A framework can break your build when it changes; this repository cannot break anything, because nothing depends on it.
That is the trade. You get zero integration cost and zero upgrade risk, and in exchange you get zero automation. A prompt library can be pinned to a commit and tested against your model; a guide page cannot. If your team needs repeatable behaviour across many repositories, the guide is an input to that work, not a substitute for it.
Within the same category of curated lists, this repository's distinguishing choice is the HowTo/Guides/Resources split. A flat awesome-style list mixes tutorials, tools and talks in one file. Separating the procedural pages from the conceptual pages from the link dumps means a reader can go straight to Guides/agents.md without scrolling past thirty tool entries. Whether the individual pages are worth the click is not something the README lets you judge.
Maintenance cost and what the repository does not settle
For a consumer, maintenance cost is close to zero: there is no dependency, no runtime and no upgrade path to track. The cost sits with the maintainer, and it is the recurring cost of auditing every outbound link and re-reading every tool page as the market shifts. A single-maintainer documentation project with a Resources branch this broad carries that cost indefinitely.
There is also a governance question the material leaves open. The licence is not stated in the README and not recorded in the repository metadata supplied here. Without a licence file, the default position under most jurisdictions is that the author retains rights, which means copying pages into an internal handbook is not obviously permitted. That is a factual gap to close by reading the repository tree, not a legal opinion, and it should be resolved before reuse rather than after.
Finally, the repository has no releases. Nothing is tagged, so there is no stable snapshot to cite internally. If you need to reference the guidance in a design document, cite a commit hash rather than the repository name.
Editorial conclusion
Adopt it as a starting index if you are a working engineer or an engineering lead who wants a single README that points at AGENTS.md conventions, coding agent skills and prompt engineering material, and you are willing to open each linked file yourself. Do not adopt it if you want something you can install, version and pin, because the repository ships no code, no schema and no release artefacts. Before relying on it, verify three things: that the linked pages under HowTo/, Guides/ and Resources/ actually exist at those paths, that the repository's licence is stated somewhere in the tree, and that the last push date is recent enough for the tool lists to still match what your team uses.
Community notes