Mohamed ElHawary

A royal index for agent workflows.

Hawary Workflow Skills is a public, MIT-licensed package for AI coding agents: nine source-backed workflows, a compatibility matrix, trust and security models, a machine-readable skills index, and Claude Code marketplace metadata. Built for repos that need structure, not a mascot costume and a motivational poster.

Live indexWorkflow Library

Nine reusable workflows now packaged with adoption docs, public trust rails, Claude metadata, and CI validation for the skill catalog. The joke is that this is still cheaper than one more alignment meeting.

Workflows
9
Agents
9 paths
Safety
Report-first

What this proves about working with me.

01I do not treat AI as magic
AI work needs boundaries, reports, verification, approval gates, and human ownership.
02I care about source-backed delivery
Useful agent workflows point to evidence instead of hiding behind confident summaries.
03I design reusable engineering systems
The repo is not a pile of prompts. It is a structured workflow library with docs and validation.
04I build for human and agent teams
The future is structured collaboration, not pretending AI writes everything responsibly by default.

Three public rails before agents touch real repos.

01Adoption Kit
Compatibility docs, machine-readable skills index, and Claude Code marketplace metadata for easier agent setup.
02Trust Kit
Trust and security docs spell out approval gates, script review expectations, read/write scope, and network behavior.
03Validation Kit
Public package tests and the skill validator now cover the index, marketplace metadata, docs links, and package hygiene.

Install or copy the skills into compatible AI agent workflows.

npx skills add mo-hawary/hawary-workflow-skills --skill <skill-name>claude plugin marketplace add mo-hawary/hawary-workflow-skills

The repo now ships the docs agents and humans need before trusting it.

01Compatibility matrix
Codex is tested; Claude Code, Claude API, Cursor, Copilot, OpenCode, Gemini CLI, Windsurf, and Cline/Roo are documented as expected paths.
02Trust model
Report-first behavior, explicit approval before edits, no hidden credential collection, and script review guidance.
03Security model
Documents read/write scope, scanner command behavior, and network calls such as OSV.dev lookups.
04Skills index
Machine-readable catalog for agent directories and package tooling.
05Claude marketplace adapter
Claude Code metadata validates locally; install command remains provisional until real marketplace smoke testing.

A compact signal deck for every workflow in the open-source repo.

01Status

Reads the room before the room becomes an incident.

Project Status Dashboard

Repo health, branch drift, docs state, and the next useful move.

Scope
Branch drift, docs state, repo health, and next action.
Relief
Gives the repo a pulse check before anyone starts prescribing vitamins.
View skill
01 / 09
Public repoAgent adoptionTrust docsCI validated

Hawary Workflow Skills

A reusable workflow library for source-backed agent work, now packaged like an adoptable public project: bilingual docs, compatibility guidance, security boundaries, Claude metadata, validation tests, and dependency pins that stop the package from aging like milk.

Open on GitHub