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 librarygit 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/library)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/library"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/library/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/library"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/library.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 | $0.00116 | $0.03515 |
| Opus 5.5 | $0.00046 | $0.01406 |
| Sonnet 5.5 | $0.00023 | $0.00703 |
| Haiku 4.5 | $0.00012 | $0.00351 |
Grade A, and why
library 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 11d 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 — 270 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ChatGPT Library
Use this as the top-level router for persistent ChatGPT Library files. Ground the target in Library, choose one content-access route, and preserve Library identity through every local edit and writeback.
The Library first-party app, connector_openai_library, owns authenticated
Library operations:
list,search,read, andfindinspect Library content.prepare_materializemakes resolved Library files available to local tools.create_library_file,replace_library_file, andmanage_librarywrite or organize Library content.sharegrants or revokes access; follow its current app-provided description and schema.- When surfaced,
prepare_uploadsandfinalize_uploadshandle prepared uploads. - App-only file
@mention search resolves selected Library items.
The runtime supplies the schemas and decides which tools are available. This skill does not expose tools or a local MCP server.
Workflow
- Ground the target.
- Decode a selected Library
@mention and reuse its identifiers. - Use
searchfor a filename, title, description, or content query. - Use
listfor recent files, folders, or inventory. - If the user supplied a local path or explicitly said the file is local, stay with local tools.
- Decode a selected Library
- Choose one access route.
- Use
readwhen Library can supply the content the task needs. - Use
findfor literal or regex matching inside known files. - Materialize when editing, scripts, visual inspection, generation, byte comparison, or another local tool needs file bytes.
- Use
- Preserve identity when writing.
- Create only when no Library identity exists or the user wants a copy.
- Replace the same
library_file_idafter editing an existing item. - Use
manage_libraryfor folders, node changes, deletion, and restore.
Use Current Helpers
- Before downloading, remove any downloaded copies of helpers from previous runs in the scoped workspace folder
- Fetch the helper and all its companion files from the current Library skill source into one new private directory.
- Reuse the downloaded helpers for every call and retry within this request or run.
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.
- 11d ago First seen · 270 lines · 116 tokens per session scan A 519e2ffb78e4
library is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (69,759 stars, last pushed yesterday), licensed CC0-1.0. It adds 116 tokens to every session and 3,515 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 2026-10-06.
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.