Works with the AI Agents you already use
Agent Production Readiness Review (Go / No-Go Scorecard)
- **Risk-scaled reviews** for read-only internal assistants, customer-facing or write-capable agents, and regulated decisions.
$9.99
Agent Production Readiness Review (Go / No-Go Scorecard)
Example session with this skill installed
Our refund agent goes live to all customers next week. It can issue refunds up to EUR 500 through a refund API, reads order history, and has a system prompt telling it to be careful. We tested it on 15 conversations manually and it looked good. There is no tracing yet. Is it ready? Write the review to readiness.md.
- Read your context and instructions
- Compiled the agent production readiness
readiness.md: verdict NO-GO. Open blockers: refunds gated only by the prompt (security.policy_gate, needs a deterministic allow/ask/deny gate), no eval set beyond 15 manual chats (quality.eval_set, quality.baseline), no tracing (ops.tracing), no evidence of budgets or a kill switch (reliability.budgets, ops.kill_switch). Conditions: idempotency keys on the refund tool, alerts with runbooks, gradual rollout. Scores by area, the remedy skill per check, and a full evidence table.
Connects securely to your tools. The creator never sees your data.
About this skill
A launch review answers one question with evidence: can this agent fail in production without hurting users, data or money, and will we know when it does? This skill turns that into a repeatable audit your agent can run on a codebase or a description.
What your agent does with it
- A weighted rubric of 32 checks in 7 areas (design and scope, tools, security and privacy, quality and evaluation, reliability, context and cost, operations), each marked blocker, major or minor, each naming the skill that fixes it.
- Evidence collection: where to look in a repository for each area, plus the anti-patterns that most often fail reviews (unbounded loops around model calls, mixed read/write tools, secrets in prompts, swallowed tool exceptions, missing stop-reason checks, prompt changes without eval runs).
- A scorecard generator (
readiness_report.py): answers template, scores per area, a verdict (GO, GO WITH CONDITIONS, NO-GO) with clear rules, open blockers and conditions with remedies, and a full evidence table in Markdown; exit codes for CI. - Risk-scaled reviews for read-only internal assistants, customer-facing or write-capable agents, and regulated decisions.
Included: SKILL.md, the checklist JSON, readiness_report.py (standard library), and an eval case (a refund agent about to go live).
Part of the Production AI Agents series. Every check points to the companion skill that remediates it; start free with Agent Architecture.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
An audit, not a fix: each failing check names the practice to apply but does not implement it. Evidence quality depends on access to the code, configs and dashboards. Regulated use cases still need legal and privacy sign-off.
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 buying a skill.
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 day ago
- Software
Python 3.10+
Permissions
File Scopes
readiness_report.py reads local JSON files and makes no network calls.