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 outlook-mailgit 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/outlook-mail)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/outlook-mail"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/outlook-mail/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/outlook-mail"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/outlook-mail.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 | $0.00022 | $0.01303 |
| Opus 5.5 | $0.00009 | $0.00521 |
| Sonnet 5.5 | $0.00004 | $0.00261 |
| Haiku 4.5 | $0.00002 | $0.00130 |
Grade A, and why
outlook_mail 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 15d 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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Outlook Mail
Purpose
Manage Outlook Mail messages with the outlook-mail companion CLI. The
connector supports both personal Microsoft accounts and Microsoft 365 work or
school accounts.
Tooling
Use exec to run:
outlook-mail <subcommand> [options]
Core subcommands:
disconnectlist [--page-size 10] [--unread] [--page-token 20] [--folder sentitems]get "<MESSAGE_ID>"search "<query>" [--folder sentitems] [--page-size 10]send --to [email protected] [--to [email protected]] [--cc [email protected]] --subject "Hello" --body "Hi ..." [--attachment /path/to/file.pdf]reply "<MESSAGE_ID>" --body "Thanks for the update!" [--reply-all]delete "<MESSAGE_ID>"(moves the message to Deleted Items; recoverable, not a permanent delete)mark-read "<MESSAGE_ID>"mark-unread "<MESSAGE_ID>"
Use --page-size for result count. --top, --limit, and --max-results
are compatibility aliases only; do not use them in new commands. Use
get "<MESSAGE_ID>" for a single message; read, get --id, and
search --query are compatibility forms only.
JSON output contract:
disconnect: parseok,action,status, anddisconnect_urllist/search: parseok,count,next_page_token(when present, pass back as--page-token),total_messages,retrieved_at, andmessages[]withid,subject,from,to,date,message_received_at,preview,is_read, andhas_attachmentsget: parseok,retrieved_at, andmessagewithid,subject,from,to,cc,date,message_received_at,body,body_type,is_read, andhas_attachmentssend: parseokandactionreply: parseok,action, andmessage_iddelete: parseok,action(trashed), andmessage_id; the moved message gets a new id, somessage_idis not the id you passed in — use the returned one for any follow-up commandmark-read/mark-unread: parseok,action, andmessage_id
Auth
Use outlook-mail --status for connector state and link management. The binary handles its callback target internally.
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.
- 15d ago First seen · 72 lines · 22 tokens per session scan A 7acb41ee9003
outlook_mail is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (69,918 stars, last pushed yesterday), licensed CC0-1.0. It adds 22 tokens to every session and 1,303 once invoked, about $0.0001 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
google-workspace
Gmail, Calendar, Drive, Docs, Sheets via gws CLI or Python.
simplify-code
Parallel 4-agent cleanup of recent code changes.
first-task
Run the first task chat that setup hands off.
initiate-setup
Run the first-run setup chat in the Hermes desktop app.
himalaya
Himalaya CLI: IMAP/SMTP email from terminal.
apple-reminders
Apple Reminders via remindctl: add, list, complete.