Platform-Specific Notes and Representative Examples
Different ecosystems create different kinds of trouble. WordPress generates more named examples than most because that ecosystem is enormous, noisy, and highly variable. Other areas are better governed by category-level rules plus selected examples.
The examples below are representative, not exhaustive.
WordPress and Related Web Stack Notes
In managed WordPress environments, the most common trouble categories include backup plugins, caching plugins, image optimization plugins, code-execution plugins, page builders, content-mutating SEO wares, and overlapping utilities that duplicate hosting or infrastructure features.
Representative examples that are commonly disallowed in our managed production and primary staging environments include:
backup wares such as All-in-One WP Migration, BackupBuddy, Duplicator, UpdraftPlus, and WPvivid
caching wares such as W3 Total Cache, WP Super Cache, WP Fastest Cache, LiteSpeed Cache, and Hummingbird
image optimization wares such as Smush, Kraken Image Optimizer, and server-based optimizers that consume local infrastructure labor
code-execution wares such as Allow PHP Execute, Insert PHP Code Snippet, and PHP Everywhere
page builders such as Divi, WPBakery, Visual Composer, Themify Builder, Brizy, and similar tools that are too difficult to optimize or govern cost-effectively
SEO or content-manipulation wares such as Broken Link Checker, Yoast, and similar tools that degrade performance, create unclear content changes, or rely on scare tactics disguised as guidance
Some exceptions do exist. For example, some builders, optimization wares, or diagnostic tools may be acceptable in qualified cases, especially in R&D, QA, migration, or other non-primary environments. Likewise, a tool may be acceptable if specific functions can be disabled cleanly and predictably.
That said, exceptions are not defaults.
Astro.js, Headless, and Front-End Adapter Notes
In headless and hybrid environments, we are cautious about wares that create unclear data flow, write directly into content structures without adequate controls, or make front-end behavior difficult to reason about.
We are especially cautious with:
brittle content connectors
weakly documented front-end adapters
tools that force excessive plugin dependencies
systems that make build behavior or deployment outcomes hard to predict
wares that create avoidable coupling between editorial systems and presentation layers
If a component makes an otherwise clean architecture harder to test, harder to migrate, or harder to hand off, it is probably not helping.
AI Platform and Tooling Notes
In AI work, the main dangers are less “plugin clutter” and more governance failure.
We avoid or tightly restrict:
ungoverned model usage
unsafe agent permissions
opaque wrappers that make prompt, output, or cost behavior hard to control
unsupported checkpoints, extensions, or model-side wares with unclear licensing or provenance
brittle retrieval and grounding setups
tools that cannot be monitored, tested, or tuned in a disciplined way
wares that make human review, rollback, or guardrail enforcement impractical
This is especially important in AI projects because bad choices here can quietly degrade quality while still looking clever in demos. We prefer boring control over exciting chaos.
Marketing, CRM, and Publishing Notes
In marketing and publishing systems, the common risk areas are fragile connectors, hidden sync behavior, unclear attribution logic, poor access governance, and automation that becomes expensive to support once it leaves the brochureware phase.
We are cautious with:
tools that sync data in opaque ways
systems that complicate deliverability, consent management, or reporting integrity
add-ons that create content or campaign changes without sufficient visibility
wares that sharply reduce portability or make account transitions painful
If a tool makes campaign operations faster in the short term but harder to govern in the medium term, that is not a bargain.
DO/IT, Access, and Infrastructure Notes
In infrastructure and operational work, disallow decisions often relate to access, recovery, logging, secrets handling, vendor sprawl, and release control.
We are especially cautious with:
unsafe access methods
unmanaged service-account patterns
weak secrets practices
tools that interfere with backup, recovery, or rollback work
wares that complicate logging, monitoring, or emergency response
components that duplicate managed infrastructure services while reducing reliability
A project can survive a lot of bad taste. It survives far less bad infrastructure.