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-skillsgit 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-skills)<a href="https://agentmods.dev/skills/asgeirtj/system_prompts_leaks/diagnosing-skills"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/diagnosing-skills/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-skills"><img src="https://agentmods.dev/badge/skills/asgeirtj/system_prompts_leaks/diagnosing-skills.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 | $0.00111 | $0.01733 |
| Opus 5.5 | $0.00044 | $0.00693 |
| Sonnet 5.5 | $0.00022 | $0.00347 |
| Haiku 4.5 | $0.00011 | $0.00173 |
Grade A, and why
diagnosing-skills 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Diagnosing Skill Configuration
Goal: reduce any skill problem to a single concrete fix. A skill "loading successfully" and a skill "triggering" are two different things — distinguish them first.
A person inspects skills in Settings → Skills and the / menu (the Skills group); an agent inspects by reading the skill files at the locations below.
1. Discovery order (earlier locations take precedence)
- Explicitly configured roots
- User
~/.zcode/skills - User
~/.agents/skills - Workspace
.zcode/skills(from the current directory up to the repository root; every level counts, and a deeper location wins) - Workspace
.agents/skills - Enabled plugin roots (lowest precedence)
Within a level, .zcode is scanned before .agents. A skill's identity is its file path, so same-named skills at different paths are all discovered, but only the first in discovery order is loaded — higher-precedence copies shadow the rest. Directories beginning with . (except .system) and node_modules are skipped.
2. SKILL.md format (two-tier failure model)
A skill is a directory containing a SKILL.md. Frontmatter is a ----delimited block of flat key: value lines (indented keys are ignored; use > or | for multi-line values). The recognized keys are name, description, when_to_use, license, and metadata.
Fails to load (the skill is dropped): frontmatter is present but name is missing, description is missing, or description exceeds 1024 characters.
Loads but may not trigger: no frontmatter at all — name defaults to the directory name and description is empty, leaving the model nothing to match on; a malformed line is skipped.
Triggering: a skill's name, description (truncated to ~250 characters), and when_to_use are presented to the model, which decides on its own whether to invoke the skill. There is no keyword matcher — a description that clearly states "when to use this" is what makes a skill discoverable to the model.
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 · 62 lines · 111 tokens per session scan A cf3ddb6d103d
diagnosing-skills is a skill published in the GitHub repository asgeirtj/system_prompts_leaks (70,207 stars, last pushed yesterday), licensed CC0-1.0. It adds 111 tokens to every session and 1,733 once invoked, about $0.0004 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.