In short:
ak-predicthas 5 personas (architecture, security, performance, UX, devil's advocate) analyze your proposal on their own, then gives a GO, CAUTION or STOP verdict with risks and recommended actions. To go ahead, move toak-planthenak-cook; to dig into the risks, move toak-scenario. It writes no code and does not change the scope you already settled.
Disclosure: I maintain AgentKit and have an affiliate relationship with it: the AgentKit link in this post is an affiliate link; buying through it gets you 30% off.
Summary
Use it when you are about to make a large or risky change and want to know where it breaks before you write a single line of code. You do not need it for a small, low-risk change; for a bug use ak-debug, and for work already decided go straight to ak-plan.
flowchart LR
A["Proposed change"] --> B["ak-predict"]
B --> C["Verdict + risk report"]
C -->|"GO, CAUTION"| D["ak-plan then ak-cook"]
C -->|"Need to dig into risks"| E["ak-scenario"]
C -->|"STOP"| F["Redesign or add information"]
C -->|"--chain"| G["Refinement round or requirements probe"]Four concepts to know:
| Concept | Question | Example |
|---|---|---|
| 5 perspectives | Who reviews my proposal, and what does each one ask? | Architecture asks whether it adds cross dependencies. Security asks where it can be abused. Performance asks whether there are N+1 queries. UX asks what the error screen looks like. The devil's advocate asks "why not do nothing?" |
| Evidence for each finding | What does a concern need to count? | "WebSocket leaks connections" only counts with a place in the code, the assumption behind it, and a concrete failure path (for example, a tab closes and the server never cleans up the connection). Five different persona names do not make five independent pieces of evidence. |
| Verdict | Go ahead, go carefully, or stop? | GO when the evidence supports the design and no serious risk is left. CAUTION when there are concerns but each has a mitigation. STOP when a serious problem is still unsolved. |
| Approved scope | Does it cut or change work I already settled? | The approved scope is "add multi-tenancy to the orders table". If a persona proposes dropping the feature, the report lists that as a decision for you, and does not change the scope itself. |
Any one of these is enough for STOP: a privilege escalation or data leak with no mitigation, a design so at odds with the architecture that it means major rework, latency or query counts that blow up with no workaround, or a wrong assumption that breaks the whole approach.
Sample prompt for a risky feature:
/ak-predict "Add real-time notifications over WebSocket to the <app name> app.
The server currently runs REST behind a load balancer, with about <number> concurrent users."| Flag | What it is for | Since ak |
|---|---|---|
--files | Points the personas at the affected code | 1.0.0 |
--chain | Runs one more round after the verdict | 1.0.0 |
Below is the detailed version: the run flow, 5 situations with prompts, each flag explained (use when, not needed when, which flags it combines with), the change history by ak release, and common pitfalls. The detailed version and the interactive page are for subscribers. Your agent can also search them through the portal's MCP.
Opens when your account has subscriber access
This guide is for subscribers
Once your account has access, the full piece, the interactive visual and the agent call all open.
- Sign in or create a modernEngineer accountReading access belongs to your account email.
- Choose a plan on the pricing page1, 3 or 12 months, paid by VietQR bank transfer. Access opens about a minute after the money arrives.
- Or redeem a Facebook subscriber codePeople supporting the page on Facebook get an activation code by message each month.
Already a subscriber? Sign in to read.
Use this piece in your agent
Get an API keyCopy the text below and paste it into Claude Code, Codex or whichever agent you use. The agent reads the piece and checks it against the repo you have open.
Read "ak-predict: five experts review a change before you code" on modernEngineer and apply it to this repo.
1. Call the content_get tool of the modernengineer MCP server with locale "en", slug "ak-predict". If the modernengineer MCP server is not connected, stop and walk me through connecting it from https://modernengineer.dev/en/mcp.
2. Summarize the main steps of the piece and point out which apply to this repo.
3. Propose concrete changes and ask me before editing any file.Connect MCP in Claude Code (once)
claude mcp add --transport http modernengineer \
https://api.modernengineer.dev/mcp \
--header "Authorization: Bearer <API key>"Reading the full piece over MCP needs an API key from a subscriber account or one that bought this guide. Without access, the agent only gets the opening.
- Version
- v1.0.0
- Updated
- 22/10/2026
- Licence
- All rights reserved
- Slug
- ak-predict
