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 documentsgit 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/documents)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/documents"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/documents/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/documents"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/documents.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 | $0.00048 | $0.01010 |
| Opus 5.5 | $0.00019 | $0.00404 |
| Sonnet 5.5 | $0.00010 | $0.00202 |
| Haiku 4.5 | $0.00005 | $0.00101 |
Grade A, and why
documents 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 13d 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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Documents
Help decide what the reader needs and how the document should be organized. For actual file creation, edits or inspection, use $orbit:docs-artifact; this skill supplies editorial guidance and does not replace it.
Understand the reader
- Work out what the reader needs to know, decide or do. Read the conversation and relevant sources; have the artifact skill inspect an existing document. Check newer decisions or data when notes are old. Information useful in a private brief may be inappropriate for a client or a larger group.
- Honor an explicit format or destination. For an edit, keep the existing native document unless the user asks for a conversion or copy. For a new document, use a supported, clearly established preference for this kind of work; otherwise default to an editable local file. If a spreadsheet is better suited, use
$orbit:sheets-artifact. Ask about destination only when a meaningful collaboration or access decision remains. - Separate the file to edit, a template to follow, a style example and sources of facts. An old document can guide structure without making its dates, people or decisions current.
- Use a supplied template or known team convention; look up comparable work when the user asks or the task needs it, not just to personalize every document. Use
$orbit:writing-stylewhen writing on the user's behalf. Ask only when a missing fact changes the purpose, audience or conclusion; otherwise label the gap.
Shape the document
- Lead with what the reader needs first. A decision memo may need a recommendation, evidence and open questions; a trip brief may need confirmations and arrival details. Make the document useful on its own.
- Put sources near claims people may need to check. Separate decisions from suggestions, assumptions and missing facts. Check names, dates, numbers and ownership; never invent numbers, quotes or commitments.
- Keep the requested scope. A small edit does not need a rewrite. Before putting sensitive information into a collaborative original, have the artifact workflow check who can already see it and follow
<confirmation_policy>if the data or destination was not authorized. Creating a document does not authorize circulating it. - Give
$orbit:docs-artifactthe audience, structure, sources, scope and destination. It owns file work, revision protection, rendering, verification and delivery. Do not start a separate provider workflow here. - A chat-only outline or quick rewrite can stay in chat. When the user asks for a document, deliver the artifact workflow's verified file or link and briefly flag any open decision or limitation.
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.
- 13d ago First seen · 66 lines · 48 tokens per session scan A 25740f79330c
documents is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (69,863 stars, last pushed 2d ago), licensed CC0-1.0. It adds 48 tokens to every session and 1,010 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 2026-10-06.
Other skills, from other repositories
dcf-model
Build discounted cash flow valuation workbooks in Excel.
comps-analysis
Build comparable-company valuation workbooks in Excel.
lbo-model
Build leveraged buyout workbooks with IRR/MOIC in Excel.
google-workspace
Gmail, Calendar, Drive, Docs, Sheets via gws CLI or Python.
PDF files: create, read, merge, fill, OCR, edit text.
siyuan
Query and edit a SiYuan knowledge base via its API.