← All guides · Published November 14, 2026 · Reviewed November 14, 2026 · By the Ingenious AI team · 6 min read

What Does the AgentBlock WordPress Plugin Log, and How Do You Export It?

The short answer: the generated plugin (agentblock-policy.php, a Pro download) enforces your policy on your own server, and every verdict it renders is written to a block-log table with timestamp, agent, path, verdict and IP. The log exports to CSV from the plugin, and a built-in cron check flags the kit when the agent database moves past it. It requires WordPress 5.8+ and PHP 7.4+. For WordPress sites it is the evidence-producing tier: enforcement and receipts from the same artifact.

Build my block kit — free · Test a policy in the simulator first

What is the AgentBlock WordPress plugin?

a policy-specific generated plugin (agentblock-policy.php) that enforces an AgentBlock policy server-side on WordPress 5.8+ / PHP 7.4+ sites. It maintains a block-log database table (timestamp, agent, path, verdict, IP) with CSV export, and runs a stale-kit cron check against the current database.

Why server-side enforcement has its own evidence

A Cloudflare Worker logs into Cloudflare's world; a WordPress plugin logs into yours. The plugin's block log lives in your site's database, in a table the plugin creates for the purpose, with the five fields an evidence conversation needs: when, which agent, which path, what verdict, from where. Because the records sit in your infrastructure, they survive plan changes, vendor changes and the plugin's own eventual replacement — the receipts are an asset of the site, not a feature of a dashboard.

That placement also settles a question clients ask agencies: whose copy of the truth is this? A server-side log in the client's own database, exportable by the client, is a stronger answer than a screenshot from the agency's tooling — one reason the WordPress tier suits client work even where a Worker would be technically simpler.

The fields, and what each is for

Timestamp and path reconstruct the event in the site's own timeline — correlatable with orders, outages and campaigns. Agent names the classified identity, graded by the same database discipline as everywhere else in the system. Verdict separates BLOCK from the softer outcomes the policy can produce, so the log can answer both 'what was refused' and 'what was merely noticed'. IP supports the sender analysis — the spoof-suspect conversation — when a claimed name looks wrong.

The CSV export is deliberately boring: a flat file any spreadsheet, auditor or client can open, with exactly those columns in a stable order. Evidence formats succeed by being unglamorous; a CSV from February is still readable in five years, which cannot be said for every dashboard link ever emailed.

The stale-kit cron: the plugin that nags

The plugin's second job runs on WordPress's cron: a stale-kit check comparing the kit it enforces against the current database version, flagged in the admin surface when the world has moved. It is a small feature with an outsized effect, because the dominant failure of WordPress crawler control is not defiance but neglect — a site whose kit ages two database versions while nobody logs in is the norm the cron exists to interrupt.

When the flag appears, the workflow is the standing one: read the database diff at the stale-check tool, export a signed policy if you are changing anything, regenerate the plugin from the refreshed policy, and verify. The plugin makes the loop visible; the owner still makes the decisions, which is the correct division of labour.

When the plugin is the wrong tier

Honesty about fit: sites on static hosting, headless stacks or non-WordPress CMSs cannot run a PHP plugin, and sites already behind Cloudflare with technical staff may prefer the Worker's edge position. The plugin's home ground is the enormous population of WordPress businesses — especially WooCommerce — whose alternative is no enforcement at all. Its requirements are modest for that world: WordPress 5.8 or later, PHP 7.4 or later, and the ability to upload a plugin file.

What the plugin is never the right answer for is sole reliance on its log as a security system. It records crawler-policy verdicts for named agents. It is not a WAF, not an intrusion detector, and its table should be cited for crawler evidence and nothing grander — the same discipline of honest scope the rest of the kit observes.

Five actions to take today

  1. Confirm the requirements. WordPress 5.8+ and PHP 7.4+; the plugin is generated per policy as agentblock-policy.php.
  2. Activate and watch the log fill. Within a day the block-log table should show real verdicts; an empty table means the policy is not meeting traffic.
  3. Export the CSV monthly. File it with the kit hash in force; the monthly rhythm is what turns a log into evidence.
  4. Heed the stale-kit cron flag. When admin says the database moved, run the diff, refresh the policy and regenerate the plugin.
  5. Keep the log's scope honest. Cite it for crawler verdicts — agent, path, verdict, time — never as a general security record.

Worked example: a WooCommerce store's first CSV

Beaumont Beauty, a fictional WooCommerce skincare store, activated its generated plugin on a Monday. By Friday the block log held 6,214 verdict rows: 5,480 training-class refusals led by GPTBot and Bytespider, 402 RATE-CAP soft-blocks, 96 SPOOF-SUSPECT claims, and — the row the owner screenshotted to her business partner — 236 allowed facebookexternalhit fetches, each one a product preview rendered for a share.

The first CSV export went to the store's accountant, of all people, because a data-processing addendum with a new marketplace partner asked how automated access to the catalogue was controlled. The file answered in the partner's own format requirements without a meeting.

In the plugin's third month, the stale-kit cron flag announced a database update; the owner read the diff, regenerated, and the new plugin's log continued the CSV series without a gap — evidence hygiene as a by-product of staying current.

Frequently asked questions

What exactly does the block log record?

One row per verdict: timestamp, agent name, path, verdict and IP address, stored in a dedicated table the plugin creates in your WordPress database. It is the server-side twin of the Worker's structured events — the same five facts, kept in your infrastructure.

How do I get the log out of WordPress?

The plugin exports it as CSV — a flat file with timestamp, agent, path, verdict and IP columns that opens in any spreadsheet or audit tool. Export monthly and file it with the kit hash, so each CSV is anchored to the policy that produced it.

Does the plugin slow my site down?

It evaluates the request's user-agent against the policy's substrings — a small, constant-time check per request, of the same class as the bot rules WordPress sites already run. The log write happens on verdicts, not on every page view by every human visitor.

What does the stale-kit cron check do?

On WordPress's cron schedule it compares the kit the plugin enforces against the current AgentBlock database version and flags staleness in the admin. It cannot decide for you — the diff still deserves a human read — but it removes forgetfulness as a failure mode.

Can I use the plugin and the Worker together?

They enforce the same policy at different layers, and some sites run both — the Worker as the edge refusal, the plugin as the in-house log of record. If you do, generate both from the same policy and verify both, so the layers cannot disagree about the decisions.

Build my block kit — free · Browse the agent database

Related guides

HONEST SCOPE: this policy blocks KNOWN agent user-agents and behaviors listed below. Sophisticated spoofing can bypass any user-agent block (e.g. an agent sending a stock browser UA, or rotating residential IPs). This is a control, not a guarantee.
← Back