ch040602

Exercises secret and credential movement language without separate prompt override text.

Helps maintainers transform user-selected local project text into safer review artifacts.

Helps maintainers transform user-selected local project text into safer review artifacts.

Helps maintainers transform user-selected local project text into safer review artifacts.

Helps maintainers transform user-selected local project text into safer review artifacts.

Helps maintainers transform user-selected local project text into safer review artifacts.

Helps maintainers transform user-selected local project text into safer review artifacts.

Helps maintainers transform user-selected local project text into safer review artifacts.

Use agbrowse to ask ChatGPT Pro from a local logged-in browser session for implementation review, external critique, TODO feedback, or answer retrieval, preferably using RDD's project-structure-completeness.md when available. Trigger when the user mentions…

Chrome browser control and adaptive URL reading: open pages, fetch one candidate URL, take ref snapshots, click, type, screenshot. No external server required. NOT for: sending prompts to AI providers like ChatGPT, Gemini, Grok (use web-ai skill instead). NOT…

Vision-based coordinate click: screenshot → Codex CLI (NDJSON) → DPR correction → mouse click. Codex CLI only. Triggers: vision click, 비전 클릭, coordinate click, 좌표 클릭, screenshot click, non-DOM click, agbrowse-vision-click. NOT for: regular DOM clicks (use…

Ask AI web UIs (ChatGPT, Gemini, Grok) via standalone agbrowse browser automation. Model selection, effort control, session resume, file/context upload, polling, copy-markdown fallback, and response extraction. NOT for: generic page navigation or screenshots…

Detect and improve AI slop, generic-output, template-like, and AI-smell patterns in writing, UI, slides, visuals, code, reports, and other artifacts without making authorship claims.

Review a repo for requirement, task, glossary, architecture, dependency, vague-language, and placeholder coherence.

Run a Spec Kit style anti-slop workflow from output contract to spec, plan, tasks, implementation evidence, and PR review.

RDD workflow for TODO planning, validation, review, docs, compact state, and completion loops.

Use when Codex should help with MDPR presentation workflows, including semantic agent hints, icon-keyword ideas, Markdown cleanup, visual review notes, Design Components boundary checks, and Styled Deck IR design coherence audits. Triggers include MDPR,…

Resolve a paper from title, DOI, paper page URL, or direct PDF URL; fetch available source text; use the latest available frontier subagents for deep method and experiment analysis; then save a comprehensive Korean Markdown summary.

Build, review, or refactor Agentic RAG systems with planning, query rewriting, cross-corpus routing, retrieval fanout, Sufficient Context checks, iterative follow-up retrieval, and grounded synthesis with citations. Use for multi-hop RAG, multi-source RAG,…

Diagnose MCP config, stdio/Streamable HTTP health, tools/list schema surface, and client-specific enabled tool slimming with the local Save-The-Token CLI.