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 diagnosing-pluginsgit 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/diagnosing-plugins)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/diagnosing-plugins"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/diagnosing-plugins/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/diagnosing-plugins"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/diagnosing-plugins.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 | $0.00131 | $0.01930 |
| Opus 5.5 | $0.00052 | $0.00772 |
| Sonnet 5.5 | $0.00026 | $0.00386 |
| Haiku 4.5 | $0.00013 | $0.00193 |
Grade A, and why
diagnosing-plugins 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 2d 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 — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Diagnosing Plugin Configuration
Goal: reduce any plugin problem to a single concrete fix.
Plugins are managed in Settings → Plugin Management — the Installed tab (enable/disable, view details, configure, uninstall) and the Discover tab (browse, install, and add marketplaces via the + button).
Key facts: enable/disable state is stored under
pluginsin~/.zcode/cli/config.json; the official marketplace iszcode-plugins-official; and for cloning marketplace repositories behind a proxy, ZCode reads the proxy fromZCODE_HTTP_PROXY(a barehttp_proxyis not used).
1. Plugin lifecycle and persistence
- Discovery sources (first match wins): inline directories, bundled official plugins, the official plugin cache, and marketplace-installed plugins. The entire subsystem is governed by the
plugins.enabledmaster switch. - Manifest location (probed in order):
.zcode-plugin/plugin.json(preferred), then.claude-plugin/plugin.json, then.codex-plugin/plugin.json. A plugin's identity is<name>@<marketplace>. - Enable/disable resolution: an explicit enable/disable entry always wins; only when no entry exists does the plugin's default-enabled status apply. A disabled plugin still appears in the list as disabled, but its components resolve to nothing.
- Persistence (all under
pluginsin~/.zcode/cli/config.json): the enable/disable map, per-plugin configuration values, and the list of suppressed (uninstalled) built-in plugins. Because a built-in plugin ships with the application and cannot be deleted, "uninstalling" one records a suppression marker that hides it from discovery. - Built-in seeding: on first launch, bundled official plugins are materialized into the plugin cache and registered in the official marketplace listing. This is idempotent and only re-materializes on a content or version change.
2. Manifest schema
- plugin.json: requires
name(matching^[a-z0-9][a-z0-9._-]{0,127}$); optionalversion(defaults to0.0.0),description,commands/skills/hooks/mcpServers, anduserConfig. - Recorded but not executed:
agents,channels,lspServers,outputStyles,settings. - Component paths are validated: an absolute path or one that escapes the plugin root is rejected as an invalid component path.
- userConfig:
typeis one ofstring,number,boolean,directory, orfile, withtitle,description,default,required, andsensitive. A sensitive value cannot currently be entered in the interface or persisted to the configuration file. - marketplace.json:
{ name, plugins[], pluginRoot?, allowCrossMarketplaceDependenciesOn? }. Eachplugins[].sourcemay be a relative path string or an object of kinddirectory,github,git,url, orgit-subdir;npmandpipare not supported.
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.
- 2d ago First seen · 60 lines · 131 tokens per session scan A 4ff859291b7a
diagnosing-plugins is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (70,207 stars, last pushed yesterday), licensed CC0-1.0. It adds 131 tokens to every session and 1,930 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 2026-10-24.
Other skills, from other repositories
ast-grep
AST-aware structural code search and rewrite via ast-grep.
systematic-debugging
4-phase root cause debugging: understand bugs before fixing.
hermes-s6-container-supervision
Modify or debug s6 services in the Hermes Docker image.
node-inspect-debugger
Debug Node.js via --inspect + Chrome DevTools Protocol CLI.
inspecting-hermes-desktop-dom
Read the live Hermes desktop DOM/CSS over CDP.
agent-merge-conflict-arbiter
Neutral arbiter for merge conflicts between two agents.