Works with the AI Agents you already use
HuG
The Humane-Gated Flow is an issue-driven development lifecycle for building software with a coding agent.
$8
HuG
Example session with this skill installed
/intake The login button is flickering on Safari, and we really need a way to export user data to CSV for the marketing team. Also, the README still shows the old API endpoint from v1.
- Read your context and instructions
- Compiled the hug
I have processed the feedback and created three distinct issues in the repository.
GitHub Issues Created
| Title | Type | URL |
|---|---|---|
| Fix login button flickering on Safari | bug | https://github.com/org/repo/issues/42 |
| Add user data export to CSV | enhancement | https://github.com/org/repo/issues/43 |
| Improve README API documentation for v2 | documentation | https://github.com/org/repo/issues/44 |
Each issue includes the original verbatim feedback as a reference. I have assigned these to you and applied the relevant labels.
Next steps
- Run
check-doc-drift README.mdto identify the specific lines in the documentation that need updating. - Use
plan-issuesto determine if the CSV export should be prioritized for the current milestone. - Start a
review-loopto begin the Safari flickering fix.
Connects securely to your tools. The creator never sees your data.
About this skill
HuG is a collection of six specialized skills designed to bridge the gap between source code and repository management. It solves the friction of maintaining documentation accuracy, managing GitHub issue hygiene, and enforcing high-quality review workflows. The package automates the tedious parts of the development lifecycle, ensuring that documentation never drifts from the implementation and that every pull request is backed by a clean, traceable history.
What it does
- Detect documentation drift by mapping repository files to specific doc sections and identifying missing or outdated content.
- Automate issue management by turning raw feedback into formatted GitHub issues with proper labels, milestones, and tracking relationships.
- Orchestrate pull requests by syncing them with existing issues, verifying CI status, and handling the merge-and-cleanup flow.
- Enforce reviewed commits through a specialized loop where the agent only commits what the human has manually staged.
- Generate repository roadmaps by sequencing open issues into pinned tracking issues with dependency logic.
How it works
- Sync check: Run
check-doc-driftagainst a file or URL to receive a diff table of the repo vs. documentation. - Intake feedback: Use
/intakewith pasted text to automatically generate categorized GitHub issues. - Plan work: Invoke
plan-issuesto build a prioritized roadmap directly within the GitHub issue tracker. - Iterative development: Use
/review-loopto write code that the human must stage before the agent will commit and push.
Frameworks & tools
This skill requires the GitHub CLI (gh) and git. It integrates with repository-specific tools like package.json scripts, Makefiles, and standard GitHub issue/PR templates.
Why this beats prompting it yourself
Writing prompts to manage GitHub API calls and multi-file drift checks is prone to hallucination and context loss. HuG provides a deterministic workflow for repository maintenance, ensuring that system-level commands like gh api and git write-tree are executed with precise, verified logic.
Use cases
- Maintaining READMEs by ensuring code snippets match the actual source files.
- Triage automation for turning messy email feedback into actionable dev tickets.
- Review enforcement for teams that want a strict "human-stages, agent-commits" workflow.
- Release planning to visualize issue dependencies without leaving the GitHub UI.
Known limitations
Requires the GitHub CLI to be authenticated. The review loop requires manual staging by the user. Planning skills are limited to approximately 200 open issues per run.
Requires authenticated GitHub CLI (gh). Review loop requires manual staging. Plan-issues limited to 200 open issues. No support for cross-repo dependencies.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 month ago
Requires
Permissions
Allowed Hosts