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-commandsgit 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-commands)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/diagnosing-commands"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/diagnosing-commands/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-commands"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/diagnosing-commands.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 | $0.00132 | $0.01772 |
| Opus 5.5 | $0.00053 | $0.00709 |
| Sonnet 5.5 | $0.00026 | $0.00354 |
| Haiku 4.5 | $0.00013 | $0.00177 |
Grade A, and why
diagnosing-commands 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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Diagnosing Command Configuration
Goal: reduce any /command problem to a single concrete fix.
A person inspects commands from the / menu (the Commands group) in the input box; an agent inspects by reading the command files at the locations below.
Two points that are often misunderstood: nested command names join with a colon (
review/code.mdbecomes/review:code, not/review/code); and a command is a.mdfile whose file name is the command name.
1. Discovery order (earlier locations take precedence)
- Explicitly configured roots
- User
~/.zcode/commands - User
~/.agents/commands - Workspace
.zcode/commands(from the current directory up to the repository root; every level counts) - Workspace
.agents/commands - Enabled plugin command roots (lowest precedence)
Within a level, .zcode is scanned before .agents. Subdirectories are scanned recursively, and a nested path joins into the command name with a colon.
2. Deduplication rule (first match wins)
The key is the normalized command name (the relative path with .md removed, separators replaced by :, lowercased). The first occurrence — the highest-precedence location — wins: user overrides workspace, .zcode overrides .agents, and local files override plugins. Duplicates are ignored.
There is also an interactive-surface-only filter: a command whose name matches a built-in slash command (or compress) is filtered out of the live / menu, though it is still present on disk.
3. Command .md format
- The command name comes from the file name and must match
^[a-z0-9][a-z0-9_:-]{0,63}$(lowercase alphanumeric start; no spaces, dots, or leading-/_; at most 64 characters). A violation drops the command. - Frontmatter (a flat parser; indented lines are ignored) recognizes
description,argument-hint,allowed-tools,model,skills, anddisable-noninteractive(all hyphenated). An unknown key is ignored but the command still loads. - A description or a non-empty body is required — otherwise the command is dropped. When
descriptionis absent, the first non-empty body line is used. - Argument substitution:
$ARGUMENTSis the full argument string;$1/$2are positional (out of range is empty). When arguments are supplied but no placeholder is present, they are appended under a "User arguments:" heading. skillsare auto-mounted. Inline dynamic shell (!`cmd`or a fenced!block) is rejected.
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 · 67 lines · 132 tokens per session scan A b11d692dbbad
diagnosing-commands is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (70,207 stars, last pushed today), licensed CC0-1.0. It adds 132 tokens to every session and 1,772 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.