Blog
News, launch notes, guides, and comparisons for builders exploring reusable AI agent skills, trust signals, and ecosystem shifts.
Latest posts
Follow product updates, editorial pieces, and practical explainers that help users navigate the marketplace and adjacent agent ecosystems.
-
Repair malformed JSON before parsing downstream automations with jsonrepair
Use jsonrepair when an agent receives JSON-like output that is almost valid but still breaks parsers, such as trailing commas, missing quotes, or concatenated fragments. The agent’s role is to normalize the payload before validation or routing, not to replace real schema design.
-
Convert DOCX documents into clean HTML for publishing workflows with Mammoth
Use Mammoth when an agent needs to turn a .docx file into simple HTML that preserves semantic structure instead of Word-specific styling. This is for ingestion and publishing workflows, not for full document editing or perfect visual fidelity.
-
Capture responsive website screenshots for visual QA with Pageres
Use Pageres when an agent needs repeatable screenshots of the same page across multiple viewport sizes for design review, release checks, or documentation snapshots. This is a bounded visual capture skill, not a general browser automation card.
-
Extract structured data and attachments from raw email with MailParser
Use MailParser when an agent receives raw RFC822 or MIME email and needs a normalized result with headers, text, HTML, addresses, and attachments. This is for email ingestion and handoff workflows, not for acting as a mailbox client or delivery platform.
-
Render git diffs as shareable HTML review reports with Diff2Html
Use Diff2Html when an agent needs to turn raw unified diffs into readable HTML artifacts for reviews, audits, or status updates. This is a rendering and packaging workflow, not a generic Git hosting, code review, or SCM platform listing.
-
Parameterize and execute notebooks as repeatable runs
Use Papermill when an agent needs to treat a Jupyter notebook like a reusable job instead of a one-off interactive document. The skill injects parameters, runs the notebook end to end, and preserves the executed output as an artifact for review or handoff.
-
Gate pull requests on OpenAPI breaking changes
Use oasdiff when an agent needs to compare old and new OpenAPI specs and decide whether a proposed change is safe to merge. The skill turns spec drift into a concrete breaking-change report that can block CI or annotate review workflows.
-
Generate adversarial API test cases from an OpenAPI or GraphQL schema
Use Schemathesis when an agent needs to turn an API schema into broad negative and edge-case coverage instead of hand-writing examples. The skill exercises live endpoints or app hooks, explores unexpected combinations, and reports failures that ordinary happy-path tests miss.
-
Run WordPress site operations safely with WP-CLI targeting and rollback guardrails
Uses the WordPress wp-wpcli-and-ops skill to guide an agent through environment-aware WP-CLI work like search-replace, plugin or theme operations, cron inspection, and multisite-safe targeting. It is an operational runbook for guarded WordPress changes, not a plain entry for the WP-CLI product.
-
Build and debug custom WordPress REST API endpoints with schema and permission guardrails
Uses the WordPress wp-rest-api skill to help an agent design, register, validate, and troubleshoot custom REST routes without skipping schema, auth, or permission_callback details. This is for agent-led endpoint work in plugins, themes, or core-adjacent codebases, not for browsing the WordPress REST API as a product catalog.