System Prompts Leaks is a collection of captured system instructions used to guide AI chatbots and coding agents before they receive user messages. It serves researchers and developers studying how different AI assistants are directed.
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 asgeirtj/system_prompts_leaks --skill a11y-debugginggit clone --depth 1 https://github.com/asgeirtj/system_prompts_leaksWrote 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/asgeirtj/system_prompts_leaks/a11y-debugging)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/a11y-debugging"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/a11y-debugging/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/asgeirtj/system_prompts_leaks/a11y-debugging"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/a11y-debugging.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 | $0.00050 | $0.01344 |
| Opus 5.5 | $0.00020 | $0.00538 |
| Sonnet 5.5 | $0.00010 | $0.00269 |
| Haiku 4.5 | $0.00005 | $0.00134 |
Grade A, and why
a11y-debugging 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 today.
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.
Copies of this mod
8 near-identical copies found in the catalogue:
- a11y-debugging — 100% identical, 19 lines differ
- a11y-debugging — 100% identical, 0 lines differ
- a11y-debugging — 97% identical, 12 lines differ
- a11y-debugging — 95% identical, 2 lines differ
- a11y-debugging — 95% identical, 2 lines differ
- a11y-debugging — 95% identical, 0 lines differ
- a11y-debugging — 95% identical, 2 lines differ
- a11y-debugging — 95% identical, 2 lines differ
How it starts
The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Core Concepts
Page Targeting: Page-scoped tools (take_snapshot, list_console_messages, evaluate_script, press_key, take_screenshot, lighthouse_audit, etc.) require a pageId parameter. Retrieve available page IDs using list_pages or from new_page.
Accessibility Tree vs DOM: Visually hiding an element (e.g., CSS opacity: 0) behaves differently for screen readers than display: none or aria-hidden="true". The take_snapshot tool returns the accessibility tree of the page, which represents what assistive technologies "see", making it the most reliable source of truth for semantic structure.
Reading web.dev documentation: If you need to research specific accessibility guidelines (like https://web.dev/articles/accessible-tap-targets), you can append .md.txt to the URL (e.g., https://web.dev/articles/accessible-tap-targets.md.txt) to fetch the clean, raw markdown version. This is much easier to read!
Workflow Patterns
1. Automated Audit (Lighthouse)
Start by running a Lighthouse accessibility audit to get a comprehensive baseline. This tool provides a high-level score and lists specific failing elements with remediation advice.
- Run the audit:
- Set
modeto"navigation"to refresh the page and capture load issues. - Set
outputDirPath(e.g.,/tmp/lh-report) to save the full JSON report.
- Set
- Analyze the Summary:
- Check
scores(0-1 scale). A score < 1 indicates violations. - Review
audits.failedcount.
- Check
- Review the Report (CRITICAL):
- Parsing: Do not read the entire file line-by-line. Use a CLI tool like
jqor a Node.js one-liner to filter for failures:# Extract failing audits with their details node -e "const r=require('./report.json'); Object.values(r.audits).filter(a=>a.score!==null && a.score<1).forEach(a=>console.log(JSON.stringify({id:a.id, title:a.title, items:a.details?.items})))" - This efficiently extracts the
selectorandsnippetof failing elements without loading the full report into context.
- Parsing: Do not read the entire file line-by-line. Use a CLI tool like
What ships with it
1 file 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.
- today First seen · 92 lines · 50 tokens per session scan A 159587fc580e
a11y-debugging is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (70,425 stars, last pushed today), licensed CC0-1.0. It adds 50 tokens to every session and 1,344 once invoked, about $0.0002 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 2026-10-30.
Other skills, from other repositories
inspecting-hermes-desktop-dom
Read the live Hermes desktop DOM/CSS over CDP.
claude-design
Design one-off HTML artifacts (landing, deck, prototype).
pretext
Build creative browser demos with DOM-free text layout.
ast-grep
AST-aware structural code search and rewrite via ast-grep.
systematic-debugging
4-phase root cause debugging: understand bugs before fixing.
sketch
Throwaway HTML mockups: 2-3 design variants to compare.