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 pwodkowski/a11y-md --skill a11y-annotategit clone --depth 1 https://github.com/pwodkowski/a11y-mdWrote 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/pwodkowski/a11y-md/a11y-annotate)<a href="https://agentmods.dev/skills/pwodkowski/a11y-md/a11y-annotate"><img src="https://agentmods.dev/badge/skills/pwodkowski/a11y-md/a11y-annotate/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/pwodkowski/a11y-md/a11y-annotate"><img src="https://agentmods.dev/badge/skills/pwodkowski/a11y-md/a11y-annotate.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00133 | $0.01923 |
| Opus 5.5 | $0.00053 | $0.00769 |
| Sonnet 5.5 | $0.00027 | $0.00385 |
| Haiku 4.5 | $0.00013 | $0.00192 |
Grade A, and why
a11y-annotate 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 2mo 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 — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.
a11y-annotate
How to read, write, and maintain a11y.md — the markdown file that captures composition-level accessibility intent for a page or prototype.
The loop
On every UI turn that touches composition (whether or not this skill is explicitly invoked):
- Read. Walk from the file being edited up to the repo root. Open every
a11y.mdfound. The union of claims is ground truth. - Apply. Make sure new code satisfies every claim relevant to the change. If the code contradicts a rationale-bearing claim the user hasn't asked to change, stop and surface it. Don't silently overwrite.
- Update. After the change, reflect any new composition-level intent in
a11y.md. Preserve everyRationale:line verbatim. Flag claim changes in the turn summary, not just the diff.
Sections
a11y.md has nine sections in fixed order. A minimal WCAG-target header sits above section 1. All sections are always present in every file; no frontmatter, no YAML, no schema.
Empty is meaningful — when a section has no claims for this page, write a single bullet saying so, with a brief reason where helpful:
## Reading order
*Empty when DOM = visual. <visual> → <DOM> for any mismatch.*
- Empty — DOM matches visual.
The bullet signals the section was considered, not forgotten.
The sections layer from wrapper to escape:
- Wrapper: Landmarks.
- Content: Headings, Images.
- Interaction: Forms, Controls.
- Consequences: Live regions.
- Reconciliation: Reading order, Focus (written last; depend on everything above).
- Escape: Notes (last resort).
When writing claims on a new route, don't attempt Reading order or Focus before the sections they integrate have been established.
Header:
**WCAG target:** WCAG 2.2 AA
Each ## header is followed — on the immediately next line, no blank in between — by a one-line italic prompt describing what belongs in the section. The prompts are part of the file format: preserved across edits, kept in empty sections as well as filled ones.
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.
- 2mo ago First seen · 182 lines · 133 tokens per session scan A 1437801d5212
a11y-annotate is a skill published in the GitHub repository pwodkowski/a11y-md (5 stars, last pushed 5mo ago), licensed MIT. It adds 133 tokens to every session and 1,923 once invoked, about $0.0005 per session on Opus 5.5. 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.
a11y-annotate compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| a11y-annotatepwodkowski/a11y-md | Skill | 5 | 5mo ago | MIT | A | 133 | Claude Code, Codex |
| penpot-mcpar27111994/penpot-mcp | Skill | 30 | 4d ago | MIT | A | 175 | Claude Code, Codex |
| antislop-humanmiqdadbadjuber/anti-slop | Skill | 8.4k | yesterday | MIT | A | 35 | Claude Code |
| critiqueplugin87/ux-ui-agent-skills | Skill | 3.1k | 10d ago | MIT | A | 37 | Claude Code |
| scaffold-projectplugin87/ux-ui-agent-skills | Skill | 3.1k | 10d ago | MIT | A | 39 | Claude Code |
| scroll-craftnateherkai/scroll-craft | Skill | 4.6k | 1mo ago | MIT | A | 177 | Claude Code |
| apply-aestheticplugin87/ux-ui-agent-skills | Skill | 3.1k | 10d ago | MIT | A | 102 | Claude Code |