You are doing a SECOND-ROUND research review for the open-source Python project glyph-arts (D:/projects/glyph-arts, github.com/2233admin/glyph-arts). v3.0.0 PyPI shipped. 5 GitHub stars. Currently: 29 chart types + ASCII art (banner pyfiglet) + image/video (chafa+ffmpeg) + Textual TUI MVP.

User has two open asks:
1. Scope upgrade: from "29 chart types" to "terminal art toolkit" — expand ASCII art, not just data viz
2. Differentiator anxiety: there are GitHub agent skills doing HTML animation. Why would users pick glyph-arts?

CODEX FIRST-ROUND CONCLUSION (your job: find what codex MISSED):

Differentiator: C+D combo. C = chart+art unified CLI. D = format flexibility (ASCII/truecolor/GIF/SVG/HTML from same input). Tagline: "terminal-native first, browser export optional".

Top wheels: borrow `art` (sepandhaghighi 2.4k), `pyfiglet`, `rich`. Wrap `chafa`, `ffmpeg`, `vhs` (charmbracelet 19k Go), `asciinema` 16k Rust, `agg` 1.5k Rust. Skip `pyxel` (game engine), `terminalizer` (Node).

Phase plan:
- Phase 2 (1-2h): `glyph-arts art TEXT --font --decor --frame --gradient` composable
- Phase 3 (2-3h): `--output txt|ansi|html-snippet`
- Phase 4 (3-4h): `glyph-arts animate` frame animation
- Phase 5 (4-5h): `glyph-arts record demo.yaml --gif` wrap asciinema/vhs/agg

"Agent skill HTML animation" — codex found NO single killer competitor, inferred category-level anxiety (vhs-action / Claude motion-skill / html-ppt-skill class).

Anti-patterns: don't chase browser motion parity; don't add Node to core; don't rewrite chafa/vhs; don't let pyxel/terminalizer enter deps; don't make Textual the main battleground; don't ship random art; don't let chart and art become disjoint.

YOUR TASK — answer 5 questions with evidence URLs (NOT echoing codex):

Q1. WHEEL SCAN GAP-FILL: Name 5+ terminal-art / ASCII-art / animation libs codex missed. Bias toward 2025-2026 new releases, Rust/Go projects, obscure-but-excellent ones, and ones with strong agent-workflow integration. Each: GitHub URL + stars + recent commit + ROI score (1-10 for glyph-arts integration).

Q2. AGENT SKILL HTML ANIMATION TRUTH: Search "claude skill HTML animation github", "agent skill terminal art animation", "MCP server animation", "claude code skill ASCII art". Find specific skill names + URLs that codex missed. Confirm: is this a category-level concern or is there a single killer?

Q3. DIFFERENTIATOR EVALUATION: Agree with C+D framing? Or disagree? Give ONE alternative framing that's MORE defensible.

Q4. PHASE 2-5 ORDERING DISPUTE: Codex order is composable-art → format-export → frame-anim → recording. Right? Which phase should jump ahead or get pushed back, and why?

Q5. ANTI-PATTERN GAP-FILL: Codex listed 8. What did codex miss? (Especially viral / community / monetization angles.)

OUTPUT FORMAT: Markdown, max 1500 words, 5 sections matching Q1-Q5, every finding with URL/repo/screenshot evidence. End with one 200-word paragraph "Three things I disagree with codex on".

CONSTRAINTS:
- Don't repeat codex's existing conclusions (that's Q0, skip)
- Don't agree with everything codex said — your value is gap-fill, not echo
- Don't write code or edit glyph-arts files
- Don't research chafa protocol (Phase 1 done)
- Stay under 1500 words

Begin research. Cite real GitHub URLs.