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 gitgit 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/git)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/git"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/git/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/git"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/git.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 | $0.00163 | $0.00927 |
| Opus 5.5 | $0.00065 | $0.00371 |
| Sonnet 5.5 | $0.00033 | $0.00185 |
| Haiku 4.5 | $0.00016 | $0.00093 |
Grade A, and why
git 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 — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Git source-control safety
The working tree belongs to the user. Make the change, leave it visible, and let them decide what becomes history.
Never without an explicit ask
The user must have asked for that exact write in this session, in their own words. An explicit request to create a new commit counts anywhere in the user's own task text, but it is not a request to amend, push, or tag. Finishing a task without that request does not authorize a history write.
commit,commit --amend,push,tag,rebase,cherry-pick,revert,merge, branch deletion,reset --hard, path restore, andclean -fall require an explicit ask.- Never publish merely because publication would be useful.
Before an authorized discard
Recovery is insurance, not authorization. A save never expands what the user allowed you to delete or overwrite.
When the user explicitly authorized an operation that may discard or broadly overwrite workspace changes, and the workspace is an ordinary Git checkout:
-
After
read_skillgives the physical Git skill package path, run:bash <git-skill-dir>/scripts/workspace-recovery.sh save before-discard -
Keep the printed
refs/tbh/recovery/...value in your handoff, then perform only the exact authorized operation. -
If the save fails or the workspace is unsupported, stop before destruction and ask the user.
To recover that saved tree later, run:
bash <git-skill-dir>/scripts/workspace-recovery.sh restore <recovery-ref>
Restore first saves the current workspace, leaves HEAD, branch, and the real
index in place, and restores the selected snapshot into the worktree. The ref
also retains distinct staged bytes when the same path had later worktree edits.
Read those staged bytes with git show '<recovery-ref>^2:<path>'; the recovery
commit's second parent is the saved index tree. Ignored files are not saved;
restore refuses an ignored-path collision rather than overwrite data it could
not save. Use only the exact printed ref, not a revision expression.
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.
- 11d ago First seen · 87 lines · 163 tokens per session scan A d7afdaf80e46
git 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 163 tokens to every session and 927 once invoked, about $0.0007 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
gh-issues
Fetch GitHub issues, spawn sub-agents to implement fixes and open PRs, then monitor and address PR review comments. Usage: /gh-issues [owner/repo] [--label bug] [--limit 5] [--milestone v1.0] [--assignee @me] [--fork user/repo] [--watch] [--interval 5] [--reviews-only] [--cron] [--dry-run] [--model glm-5]…
caveman-commit
Write a Conventional Commits message compressed to intent only. Use for "write a commit", "commit message", /commit or /caveman-commit.
git-advanced-workflows
Master advanced Git workflows including rebasing, cherry-picking, bisect, worktrees, and reflog to maintain clean history and recover from any situation. Use when managing complex Git histories, collaborating on feature branches, or troubleshooting repository issues.
agent-merge-conflict-arbiter
Neutral arbiter for merge conflicts between two agents.
github
GitHub via gh CLI: PRs, issues, reviews, repos, auth.
changelog
Auto-generate a changelog from git commits and sprint data. Internal and player-facing versions.