Reads exported logs
Counts requests per address and per path after a spike, using your own log-summary script.
Product
Stop DDoS attacks? Plan the layers you run yourself
01
Nobody can promise to stop DDoS attacks with one app, and SmartCode Relay does not try. It sees no traffic. It reads the logs you export from your own servers and sites, drafts rate limits and a response runbook, and shows file changes as diffs for you to review and test.
Drafts rules and runbooks
Writes rate-limit drafts and a plain-text response runbook; in Ask first mode you see the diff before you accept.
Stays outside traffic
It works on files after the fact and has no place in your traffic path.
What the agent can do
- Filesread_file opens your settings and exported logs, grep and glob search them, and edit_file and write_file draft changes and the runbook in the chat's working folder.
- Shellruns your own log-summary script and configuration test command, then reports the real exit code. Dangerous commands are blocked.
- Webweb_search and http_get read the documentation of your hosting provider or web server, so setting names come from the source.
- Screenshots and imagesocr reads text from a screenshot of a provider dashboard or an error page, and gen_image saves image files.
- Orchestrationtask runs sub-agents in parallel, run_on_model sends a hard step to a bigger model, and skill keeps reusable playbooks such as your runbook format.
- Approval modesAuto-run acts at once, Ask first pauses before actions that change something, and Plan only writes a plan and runs nothing until you switch mode.
- Cronruns a prompt on a schedule, for example the same log review on the exports you drop into the project folder each morning.
- Modelsuse a cloud AI provider through a standard API endpoint, a model running locally on your PC, or the built-in model.
Where the agent fits in a layered defence
Preventing DDoS attacks outright is not realistic. Capacity sits at your provider. At the application level you cannot prevent a DDoS flood alone, so the app helps you draft per-client limits and document the layers.
Read the log summary, draft a rate limit for the busiest path and test it on the staging copy.
- read_file
exports/summary.csvread - grep
/checkout summary.csv3 matches - read_file
staging/server-settings.txtread - edit_file
staging/server-settings.txt+6 -1 - Shell
your test commandexit 0 - write_file
notes/rate-limit-change.txtsaved
Waiting for your OK before the next stepRejectAccept
Two tasks, step by step
Read the numbers behind a spike
Telling traffic spikes from layer 7 DDoS floods by hand takes three sources side by side. read_file opens the exported analytics report and your campaign calendar, Shell runs your log-summary script, and grep searches the saved summary for the paths you suspect. A launch brings many networks and varied pages, while a flood from many sources often shows identical requests to one or a few paths at very high rates per client. Logs give hints, not proof. The agent only summarises counts per address and per path, and you make the judgement. Compare the summary with your own analytics, and ask your hosting provider when the picture is unclear.
Check your own scripts for retry loops
A runaway script, a retry loop or a misconfigured monitor can overload your own server by accident. To find that, point grep at your scripts and monitor settings for retry and interval lines. read_file opens each hit, and edit_file proposes a longer delay with a diff. Shell then runs your test command on a copy and reports the exit code. Internal clients need limits too. If your home internet is getting DDoS traffic instead, what to do first is call your internet provider. In Ask first mode every edit pauses for your OK, so read the diff and run the test again.
What SmartCode Relay does not do
It runs on your PC. SmartCode Relay is not a hosted service or a plugin, and it is meant only for your own files and servers you are authorised to use. It does not enforce that scope, so you keep it away from anything you do not own. Nothing here guarantees correct code, scans or analysis.
