Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add philmingdao/anno --skill review-html-artifactsgit clone --depth 1 https://github.com/philmingdao/annoWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/philmingdao/anno/review-html-artifacts)<a href="https://agentmods.dev/skills/philmingdao/anno/review-html-artifacts"><img src="https://agentmods.dev/badge/skills/philmingdao/anno/review-html-artifacts.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00071 | $0.00887 |
| Opus 5 | $0.00036 | $0.00443 |
| Sonnet 5 | $0.00014 | $0.00177 |
| Haiku 4.5 | $0.00007 | $0.00089 |
Grade A, and why
review-html-artifacts scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 38 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review HTML Artifacts
Use the anno plugin's html_review_* MCP tools for session state and file generation. Use the host's browser capability when available; otherwise return the local review URL so the user can open it.
Start a review
- Call
html_review_start_sessionwith the absolute source HTML path. Omitoutput_pathunless the user names a destination. Never overwrite the source file. - Open the returned
review_urlwith the available browser capability. If the host cannot open URLs, give the URL to the user without claiming it was opened. - Tell the user they can edit text in place, use the compact format toolbar, drag a selected element by its top handle, add an element comment from the lower-left comment button, or drag a rectangle from blank space to annotate a region. The page autosaves all of this to the MCP session.
- Let the user click 提交给 Agent. Anno writes a clean draft containing direct text, formatting, and visual-position edits, stores that draft plus every element annotation, area annotation, and page note, and marks the session
needs_agent. - Anno persists a generation-specific handoff before attempting host delivery. Hosts with a conversation bridge may send a follow-up automatically. In other hosts, tell the user to return to the agent and continue with the session id. Never claim that a follow-up was delivered without a handoff receipt.
- On receiving a handoff, call
html_review_claim_handoffimmediately with the exactsession_idandhandoff_idfrom the message. This durable receipt is what allows Anno to show that the Agent received the request. Then reload the session and take ownership of generation, validation, registration, and user-facing delivery.
Finish a review
- Call
html_review_claim_handofffirst when the follow-up contains a handoff id, then callhtml_review_get_session. Claiming is idempotent for the active handoff. - For a
needs_agentsession (or a legacyneeds_codexsession), usesession.draft_html_pathas the only HTML base. Preserve direct edits and format/position changes already present in the draft, then apply every unresolved element annotation, area annotation, and page note from the review payload. - Write a complete standalone HTML file to a temporary non-source path, preserve scripts/assets/fonts/doctype, remove Anno-only review attributes, and validate the first page plus every affected page. For ordinary pages, validate the edited viewport and responsive behavior.
- Call
html_review_register_finalwith the verified file. This copies it to a versioned non-overwriting destination and marks the sessionresolved. - If a legacy session is still
generating, wait for it to finish instead of duplicating that old worker. If it hasgeneration_error, report it and recover only after correcting the underlying problem. - Return the final absolute path and summarize applied annotations. Do not claim free-form notes were applied unless the registered output was checked.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 38 lines · 71 tokens per session scan A bc2cc79c6d66
review-html-artifacts is a skill published in the GitHub repository philmingdao/anno (1 stars, last pushed 17d ago), licensed MIT. It adds 71 tokens to every session and 887 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
annotation-feedback-wait
A workflow for collecting and waiting for visual feedback that users draw on a web page with a browser annotation tool.
popular-web-designs
54 real design systems (Stripe, Linear, Vercel) as HTML/CSS.
figma
Import Figma content into a HyperFrames composition — rendered assets, brand tokens, components, storyboard sections → reconstructed motion (frames read as states, not slides) (REST/CLI), connector-assisted motion when available, and shaders from a connector or native export. Use when the user pastes a figma.com link…
motion-graphics
A short, design-led motion graphic where motion is the message — kinetic typography, stat count-up, chart/data-viz hit, logo sting / brand lockup, lower-third / callout / social overlay, animated map (highlight regions, connect places, zoom to a location), animated tweet / news-article / headline, webpage / UI…
figma-to-webui
Recreate Figma design frames as production-ready Chromium WebUI code following Chrome's Lit-element architecture and styling guidelines. Use when asked to create WebUI code from a Figma frame URL.
generate-component-spec
Maps a Figma component from the Chrome Design System (CDS) to equivalent C++ Views, WebUI, and Clank (Android) components in the Chromium codebase. Generates a comprehensive side-by-side comparative Markdown specification detailing layout variants, interactive states, design tokens, architectural and implementation…