deepwork
Repeatable multi-step workflows with automated quality reviews
All Skills in This Repository (24)
Pure Emerald Level Indicatorsmanual_tests.infinite_block_tests
Run four infinite-block tests in strict order with sub-agents.
deepwork_rules.define
Create deterministic file-change governance rules with YAML frontmatter.
commit
Automate code review, tests, linting, and commit-and-push workflows.
deepwork_jobs.define
Create DeepWork job.yml specifications through interactive workflow requirement questions.
commit.lint
Format and lint Python code with ruff after tests pass.
commit.commit_and_push
Commit and push validated code changes after lint passes.
deepwork_jobs.learn
Extract learnings from DeepWork job conversations and update instruction files.
add_platform
Generate adapters, templates, and tests for new AI platforms in DeepWork.
commit.test
Pull latest code and run the project's test suite until all tests pass.
add_platform.verify
Verify end-to-end installation and integration of a new AI platform in DeepWork.
update
Updates Deepwork standard jobs and syncs them via install command.
commit.review
Review changed files for quality issues, DRY violations, naming clarity, and missing tests.
Frequently Asked Questions
FAQPage SchemaHow to install DeepWork?▼
Run `npx skills add Unsupervisedcom/deepwork --all -g -y` in your terminal to install all skills in this suite globally.
How to make AI agents reliable on complex tasks?▼
DeepWork forces the agent through structured step-by-step workflows with quality gates, so it cannot skip steps or go off-script. Automated reviews then verify every output against your rules.
How to turn a one-off task into a repeatable workflow?▼
Do the task once with your agent, then ask DeepWork to capture it as a job. It generates a hardened multi-step workflow you can rerun anytime with a single command.
Does DeepWork work with Claude Code and Gemini CLI?▼
Yes. DeepWork ships as a plugin for Claude Code and an extension for Gemini CLI, with an MCP server as the shared runtime.
Can non-developers use DeepWork workflows?▼
Yes. Workflows are domain-agnostic, so research, reporting, and marketing tasks work the same way as coding jobs, all driven by plain-English requests.
Related Repositories in Software Engineering
View All in Software Engineering→openclaw
Run a personal AI assistant across your devices and chat apps
superpowers
Gives coding agents a disciplined workflow from idea to merged code
react
AI agent skills for building, testing, and porting React core