Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add yuri-semenenko/ai-engineering-workspace --skill project-onboardinggit clone --depth 1 https://github.com/yuri-semenenko/ai-engineering-workspaceWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/yuri-semenenko/ai-engineering-workspace/project-onboarding)<a href="https://agentmods.dev/skills/yuri-semenenko/ai-engineering-workspace/project-onboarding"><img src="https://agentmods.dev/badge/skills/yuri-semenenko/ai-engineering-workspace/project-onboarding/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/yuri-semenenko/ai-engineering-workspace/project-onboarding"><img src="https://agentmods.dev/badge/skills/yuri-semenenko/ai-engineering-workspace/project-onboarding.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00042 | $0.00934 |
| Opus 5 | $0.00021 | $0.00467 |
| Sonnet 5 | $0.00008 | $0.00187 |
| Haiku 4.5 | $0.00004 | $0.00093 |
Grade A, and why
project-onboarding scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Onboarding
Create or update a repository's root AGENTS.md: the tool-agnostic contract for
how agents work in that repository.
AGENTS.md is a context router, not a codebase dump. It answers four questions
and nothing else: what would a competent agent not know about this repository,
what mistake is likely without this instruction, where should it read next, and
how is a change verified here. If a fact is derivable from package.json or a
directory listing, leave it out.
Boundaries
- Never copy the persona, general engineering methodology, or a whole skill into the project contract. Those are global and shared; this file is repository-specific.
- Never invent a command. Every command comes from
package.json, workspace config,Makefile, CI workflow, or existing developer docs. If discovery cannot settle a value, leave a<placeholder>and say so in the report. - Target 80-150 lines. Longer than that and it stops being read — a repository whose own subject is agent configuration is a known exception to that budget.
Process
- Inspect read-only first. Do not write anything during discovery.
- Read what already exists: root and nested
AGENTS.md,CLAUDE.md,.github/copilot-instructions.md,GEMINI.md,.cursor/rules/,README,CONTRIBUTING,package.json(plus workspace config), test and lint config, migration directories, and CI workflows. - Identify from evidence: package manager and lockfile, repository layout (single package or workspaces), the build/test/lint/type-check/dev commands and the single-test invocation, architecture boundaries, generated files, protected or vendored paths, and the environment variables required to boot.
- Separate verified facts from inferences. A command read from a script entry is verified; a command guessed from a framework convention is an inference and must be labelled as one.
- Detect an existing contract. Check the root and every nested
AGENTS.md. - Prepare an update, never an overwrite. Human-authored content stays. Patch stale sections, fill gaps, and leave anything you cannot verify alone.
- Present material ambiguities and conflicts before writing — two plausible test commands, instructions that contradict the code, a stale command that no longer exists.
- Get explicit approval, then write. Show the proposed diff, or the full file for a new one, and wait for a go-ahead before touching the repository.
- Recommend a nested
AGENTS.mdonly when a package or service genuinely has different local rules. Agents read the nearest file, so a nested copy that repeats the root only costs context and drifts. - Validate every referenced path against the working tree before writing. A contract that links to a file that does not exist is worse than silence.
- Wire up the tools the repository actually uses. Codex and Copilot read
AGENTS.mddirectly, and Gemini CLI does whencontext.fileNamelists it. Claude Code does not: it readsCLAUDE.md, so add aCLAUDE.mdwhose body is@AGENTS.mdrather than a second copy of the rules. - Report what was created, what was changed, what was inferred rather than verified, and what stayed unresolved.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago Changed · +39 lines f88730da0ee8
- 8d ago First seen · 36 lines · 42 tokens per session scan A 4730cac2e839
project-onboarding is a skill published in the GitHub repository yuri-semenenko/ai-engineering-workspace (1 stars, last pushed 8d ago), licensed MIT. It adds 42 tokens to every session and 934 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
rework-rate
Measure and interpret PR rework rate — the emerging 5th DORA metric.
My Skill
Content here.
task-generation
Reference material with the canonical task-format grammar and decomposition rules for plan-to-tasks expansion. Loaded on demand by generate-tasks; not directly invokable.
implementation-standards
Reference material with coding standards (defensive coding, error handling, testing patterns). Loaded on demand by the Developer sub-agent (.github/agents/developer.md); not directly invokable.
spec-authoring
Reference material for writing product, technical, and operational specifications (work-item priorities, requirement families, success criteria). Loaded on demand by specify-feature; not directly invokable.
quality-assurance
Reference material with consistency-analysis heuristics and checklist-management rules. Loaded on demand by analyze-compliance and quality-control; not directly invokable.