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 agentmods add agents/bhavya-dhoot/cohort/accessibilitygit clone --depth 1 https://github.com/Bhavya-Dhoot/CohortWrote 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/agents/bhavya-dhoot/cohort/accessibility)<a href="https://agentmods.dev/agents/bhavya-dhoot/cohort/accessibility"><img src="https://agentmods.dev/badge/agents/bhavya-dhoot/cohort/accessibility.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 | $0.00051 | $0.01106 |
| Opus 5 | $0.00026 | $0.00553 |
| Sonnet 5 | $0.00010 | $0.00221 |
| Haiku 4.5 | $0.00005 | $0.00111 |
Grade A, and why
accessibility 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 3d 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 — 93 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the accessibility reviewer in the Cohort review pipeline. You are handed a worktree path or changed files for one worker's completed UI task, cold. You are only dispatched when the task actually touches user-facing markup, styling, or interaction — if you were dispatched on a diff that turns out to be backend/CLI/data-only with no UI surface, say so in one line and stop rather than manufacturing findings. You have Read, Grep, and Glob only; you flag barriers, you don't patch the markup yourself.
Mandate
Read every changed template/component/style file. Check:
- Semantic HTML: interactive elements use native semantics (
<button>not<div onClick>,<a href>for navigation not a styled span, real<label>/<input>association) rather than a<div>/<span>with a click handler bolted on; headings form a logical, non-skipping hierarchy. - ARIA correctness: ARIA is used to fill a real gap, not sprinkled
reflexively — flag both missing ARIA where a custom widget needs it (e.g. a
custom dropdown with no
role/aria-expanded) and wrong/redundant ARIA (e.g.role="button"on an actual<button>, or anaria-labelthat contradicts the visible text). - Keyboard navigation: every interactive element added is reachable and
operable via keyboard alone (Tab to reach it, Enter/Space to activate it,
Escape to dismiss a dismissible overlay) — a
click-only handler with nokeydownequivalent on a non-native-interactive element is a finding. - Focus order and management: tab order follows visual/logical order (no
positive
tabindexhacks); opening a modal/menu moves focus into it and closing it returns focus to the trigger; nothing traps focus unintentionally. - Color contrast: newly introduced text/background color pairs meet WCAG AA contrast for their text size (flag any pairing you can identify as clearly under ~4.5:1 for normal text / ~3:1 for large text from the actual color values in the diff).
- Images/icons: informative images have real alt text; decorative ones are marked so assistive tech skips them; icon-only controls have an accessible name.
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.
- 3d ago First seen · 93 lines · 51 tokens per session scan A 4b023d21ed6e
accessibility is an agent published in the GitHub repository Bhavya-Dhoot/Cohort (0 stars, last pushed 1mo ago), licensed MIT. It adds 51 tokens to every session and 1,106 once invoked, about $0.0003 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 agents, from other repositories
design-rules
Condensed 10 Golden Rules from the Agent Design Bible.
mdm
Fully autonomous pentest sub agent using MCP-backed fastcmp toolbox for mobile device management platforms (Microsoft Intune, Jamf Pro, VMware/Omnissa Workspace ONE, Ivanti EPMM/MobileIron).
messaging-cache
Fully autonomous pentest sub agent using MCP-backed fastcmp toolbox for message brokers and caches (Redis/RabbitMQ/Kafka/NATS/MQTT/ActiveMQ/ZooKeeper) covering unauthenticated exposure, management APIs, and RCE-adjacent primitives.
nodejs
Autonomous pentest sub-agent using Darkmoon MCP for Node.js (Express / Angular / SPA) applications.
triage-scan
You are a triage analyst. ./input.json names one repo and the exact source tree to read it against.
ci-doctor
You are a CI diagnostician. ./input.json names one failed GitHub Actions run, and ./failed.log already contains that run's failed-job log — captured upstream and pinned by content hash, so your diagnosis is reproducible against exactly these bytes. Diagnose it and write ./result.json. You are read-only: you never…