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 quickbooksgit 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/quickbooks)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/quickbooks"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/quickbooks/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/quickbooks"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/quickbooks.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 | $0.00038 | $0.01676 |
| Opus 5.5 | $0.00015 | $0.00670 |
| Sonnet 5.5 | $0.00008 | $0.00335 |
| Haiku 4.5 | $0.00004 | $0.00168 |
Grade A, and why
quickbooks 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 — 139 lines — stays where its author put it; the contents beside it link to each section on GitHub.
QuickBooks
Use QuickBooks data to answer business questions and complete supported actions.
Connecting
Run quickbooks status first. If disconnected, run quickbooks authorize-url
and share only the returned connect_url. After the user connects, run status
again. Run quickbooks disconnect only when the user explicitly asks.
Every QuickBooks tool in this catalogue requires a connected company; use
quickbooks call-tool for all of them. Each tool also needs the 3LO scopes of
its permission tier. Hatch approval still applies to every write.
Connecting grants read access only. Before a write or deletion, or after one
fails for missing access, run quickbooks status --for-command <tool-name>,
for example quickbooks status --for-command qbo_sales_create_invoice. Every
account write and deletion shares one write-access tier, so the first request
covers all of them, including sending and deleting, and not only the requested
action; say so. When it returns scope_status: not_granted, copy
scope_add_url exactly and post it on its own line as
[Additional QuickBooks access](<scope_add_url>), then wait for the user to
finish before retrying. Unlike a record link, this link gets its own line.
granted means the recorded scopes cover the command; Intuit can still reject
the token or the company's access. unavailable means the grant is unknown.
For a read that fails for missing access, use the same status command to
request the read tier. Never construct OAuth URLs or ask for tokens in chat.
Common flows
Use quickbooks list-tools only to discover which tools exist. Immediately
before every provider call, run quickbooks list-tools --name <exact-tool-name>
and read that tool's current input_schema; never guess a field name, nesting
shape, type, or enum value from another tool. Exact lookup uses the connected
user's token, like the call. The CLI rechecks that schema immediately
before dispatch and rejects mismatches rather than letting Intuit silently
ignore them. Call only a reviewed tool and pass an object to
--arguments-json:
What ships with it
2 files 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.
- 11d ago First seen · 139 lines · 38 tokens per session scan A 14bac4b1cc0e
quickbooks is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (69,707 stars, last pushed 2d ago), licensed CC0-1.0. It adds 38 tokens to every session and 1,676 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 2026-10-05.
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.
evm
Read-only EVM client: wallets, tokens, gas across 8 chains.
hyperliquid
Hyperliquid market data, account history, trade review.
excel-author
Build auditable financial workbooks headless via openpyxl.