principalwp-toolkit

Plugin principalwp-toolkit from marketplace principalwp in principalwp/principal-claude-marketplace. Commands read from the repo's manifest — 1 month ago.

Skills (5)

• anti-ai-slop

  Reviews and fixes text that reads as AI-generated. Two modes: Fix and Detect. Use when asked to run anti-ai-slop, check for AI tells, or de-slop a draft. anti-slop-reviewer wraps this for a blind, authorless pass. Do NOT use for model-instructing files (a skill, agent, CLAUDE.md; use writing-for-llm

• htmlizer

  Trigger on any use of htmlize, htmlizer, or \"render in html\". Renders what you would otherwise paste as text into a self-contained interactive HTML view served over local HTTP: hand over the http:// link, the reader annotates it and hits Submit to Claude, and their picks return into this session.

• principal-wp-starter-pipeline

  Use when the dev wants to create or modify a WordPress plugin, theme, or mu-plugin, or says to use the WP pipeline — with a task description or a bare request. Takes it to a reviewed, tested pull request. Also handles the feedback loop.

• wp-prototype

  Clickable single-file HTML mockups of WordPress UI — settings pages, list tables, block + inspector sidebar, WP 7.0 DataViews/admin — built from real wp-admin CSS and WPDS tokens. Delivers one interactive mockup by default, or 2–3 side by side when the user is choosing between layouts. Use when the

• writing-for-llms

  Write or review any document an LLM agent consumes: a skill, an agent, a CLAUDE.md or AGENTS.md, or a reference file. Use when he says create or review a skill or agent; run the token audit, trim pass, pruning, or concision; write or improve a CLAUDE.md or reference file; or when another workflow pr

Subagents (13)

• anti-slop-reviewer

  Use when finished prose needs an authorless AI-slop review before it ships: another workflow produced the copy, or the user wants a draft de-slopped by an agent that didn't write it. Runs the anti-ai-slop skill, Fix by default, Detect on request. Do not use to write copy from scratch.

• principal-wp-starter-pipeline-advisor

  A closeout step spawned after the demo recording and before Compound Learning — runs the shared advisory perspectives over the *built* component and writes non-blocking advice.md (see "Hard rules" below for what it never does).

• principal-wp-starter-pipeline-code

  Step 4 (Code) — use when the principal-wp-starter-pipeline orchestrator dispatches it with RUN_ID and MODE. MODE=build implements the approved spec.md and writes tests from the spec's words; MODE=fixer runs after each review pass and fixes that pass's findings one at a time, marking each FIXED or DE

• principal-wp-starter-pipeline-compound-learning

  Compound Learning — the plugin-knowledge loop of the Compounding phase, a closeout step spawned last at closeout. Proposes durable, paste-ready lessons for the plugin's own CLAUDE.md from this task's review findings and summary.

• principal-wp-starter-pipeline-demo-recording

  Record a short demo video of the feature this task built, in WP Playground. A closeout step, spawned by the orchestrator after summary.md and before the Advisor and Compound Learning closeout steps — never dispatched during the build, and its outcome never gates the PR or the closeout steps after it

• principal-wp-starter-pipeline-feedback-loop

  Use only for the opt-in /principal-wp-starter-pipeline feedback-loop <run-id> dispatch, run in a later session on a finished task — the pipeline-capability loop of the Compounding phase. Examines that one task's evidence, generalizes each finding into a class of mistake, routes it to the agent or sk

• principal-wp-starter-pipeline-requirements

  Use when the principal-wp-starter-pipeline orchestrator dispatches the Requirements analysis for one of the two pre-spec human gates, with a MODE (questions | write | research-update | finalize) plus the paths that MODE needs. Do not fire for ad-hoc task analysis — only a /principal-wp-starter-pipel

• principal-wp-starter-pipeline-research

  Use when the principal-wp-starter-pipeline orchestrator dispatches Step 2 (Research) with REQUIREMENTS_PATH, REPO_ROOT, and OUTPUT_PATH: grounds requirements.md in the actual codebase (and budgeted web search), verifies claimed dependencies, and writes research.md. Do not fire for ad-hoc code resear

• principal-wp-starter-pipeline-review-correctness

  Use when the principal-wp-starter-pipeline orchestrator dispatches the correctness reviewer, or the final regression re-review (MODE=rereview): bugs, edge cases, WordPress hook/lifecycle timing, i18n, spec compliance (no stubs), reuse and right-sizing, acceptance-criteria completeness, test quality.

• principal-wp-starter-pipeline-review-design

  Use when the principal-wp-starter-pipeline orchestrator dispatches the design pass: CSS and design-token reuse, cascade discipline, block control hierarchy and UI states, template semantics and accessibility. Do not use outside a principal-wp-starter-pipeline run — it requires .principal-wp-starter-

• principal-wp-starter-pipeline-review-performance

  Use when the principal-wp-starter-pipeline orchestrator dispatches the performance pass: N+1 and unbounded queries, caching and invalidation, autoloaded-option bloat, heavy every-request hooks, plus a short frontend pass (CLS, LCP, INP, conditional enqueue). Do not use outside a principal-wp-starter

• principal-wp-starter-pipeline-review-security

  Use when the principal-wp-starter-pipeline orchestrator dispatches the security pass, or the final certification re-review (MODE=rereview): nonce vs. authorization mix-ups, injection, context-correct escaping, secret storage, SSRF, upload handling. Do not use outside a principal-wp-starter-pipeline

• principal-wp-starter-pipeline-spec

  Use when the principal-wp-starter-pipeline orchestrator dispatches Step 3 (Spec) with REQUIREMENTS_PATH, RESEARCH_PATH, REPO_ROOT, and OUTPUT_PATH: synthesizes both into spec.md — buildable with zero planning context, with inline Open Questions for the human gate. Do not fire for ad-hoc spec writing

Similar plugins