- Home
- Skills
- Agents & Orchestration
- Human Review Publish
Works with the AI Agents you already use
Human Review Publish
The one explicit checkpoint where a human — not the agent — decides a change is ready to go public. No commit, push, or PR without a real, current, specific approval.
$4.99
Human Review Publish
Example session with this skill installed
The acceptance criteria all passed for PROJ-109 — show me what's ready to ship.
- Read your context and instructions
- Compiled the human review publish
Summary
Presented the review checkpoint — nothing pushed yet. After explicit approval ("yes, push and merge"), published for real: branch pushed, PR opened, CI passed, merged.
| Phase | Result |
|---|---|
| Review (always first) | 3 files changed, all checks passed, AC complete — awaiting approval |
| Approval given | "yes, push and merge" |
| Publish | Pushed, PR opened, CI 9/9 green, merged |
{
"phase": "published",
"changedFiles": [".claude/skills/acceptance-criteria-verification/SKILL.md", "...2 more"],
"checksRun": ["check_skill_conformance.py: passed", "semgrep: 0 findings", "gitleaks: no leaks found"],
"acceptanceCriteriaStatus": "complete",
"publicationPerformed": true,
"commitSha": "68e8dad",
"branchName": "PROJ-109-acceptance-criteria-verification-skill",
"prUrl": "https://github.com/your-org/your-repo/pull/72"
}
Next steps
- Nothing — this ticket is fully shipped. The next ticket starts a fresh review; this approval never carries forward to it.
Connects securely to your tools. The creator never sees your data.
About this skill
Every other skill in this suite stays read-only or local-only. This is the one place the chain can reach something externally visible — and it only gets there after a real, current, specific human approval, never an assumption that a change "obviously" looks fine.
It always stops first: assembles a full review summary (changed files, every check run, acceptance-criteria status, known risks) and states plainly that nothing has been published yet. Only on a separate, later invocation — with approval naming this exact reviewed change set — does it commit, push, and open a PR, mirroring the real verification results, never a fresh unverified claim.
Its most important rule: an approval never carries forward. If the code changes after approval — even a small, obviously-safe follow-up — the old approval doesn't cover it. A fresh review gets presented, every time. No force-push, ever, under any approval.
Part of the Ticket-to-Shipped workflow: 10 skills covering the full requirement-to-shipped-code lifecycle. Works standalone too.
Needs an actual human in the loop to approve — this skill will never invent or infer an approval. Doesn't work around a repository with no PR/pull-request concept (a trunk-based push-only workflow would need adapting). If code changes after approval, it always asks again — by design, so expect a fresh review on every edit, however small.
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 18 days ago
Requires
Permissions
Never force-pushes or performs any destructive action — strictly commit, push, and PR creation, and only after explicit approval naming the exact change set.