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.
git clone --depth 1 https://github.com/ondrej-svec/heart-of-gold-toolkitWrote 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/plugins/ondrej-svec/heart-of-gold-toolkit/marvin)<a href="https://agentmods.dev/plugins/ondrej-svec/heart-of-gold-toolkit/marvin"><img src="https://agentmods.dev/badge/plugins/ondrej-svec/heart-of-gold-toolkit/marvin/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/plugins/ondrej-svec/heart-of-gold-toolkit/marvin"><img src="https://agentmods.dev/badge/plugins/ondrej-svec/heart-of-gold-toolkit/marvin.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
marvin 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 10d 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.
What it actually says
{
"name": "marvin",
"version": "0.7.0",
"description": "The Paranoid Android \u2014 quality tools for code review, knowledge compounding, work execution, and grounding",
"author": {
"name": "ondrej-svec",
"url": "https://github.com/ondrej-svec"
},
"homepage": "https://github.com/ondrej-svec/heart-of-gold-toolkit/tree/main/plugins/marvin",
"repository": "https://github.com/ondrej-svec/heart-of-gold-toolkit",
"license": "MIT"
}What it installs
The manifest is a name and a version. 13 skills, 3 agents, 2 hooks travel with it, and installing the plugin installs all of them — 1,277 tokens a session between them. Each is measured on its own page, and each can be installed alone.
- Skill copy-editor A 202 tokens
- Skill ground A 116 tokens
- Skill work A 71 tokens
- Skill compound A 78 tokens
- Skill redteam A 61 tokens
- Skill scaffold A 51 tokens
- Skill test-writer A 68 tokens
- Skill quick-review A 59 tokens
- Skill share-server-setup A 53 tokens
- Skill share-server-control A 44 tokens
- Skill share-html A 52 tokens
- Skill harness-up A 112 tokens
- Skill harness-sync A 109 tokens
- Agent harness-author A 103 tokens
- Agent skill-reviewer A 57 tokens
- Agent knowledge-architect A 41 tokens
- Hook PostToolUse A not measured
- Hook PreToolUse A not measured
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.
- 10d ago First seen · 12 lines scan A 54515cb8565c
marvin is a plugin published in the GitHub repository ondrej-svec/heart-of-gold-toolkit (19 stars, last pushed 20d ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 2026-08-30.
Other plugins, from other repositories
session-checkpoint
Session closer — saves context before compact. Handoff file, memory updates, lesson extraction, reflexion (what went wrong, what to improve).
doc-drift
Audits the memory and documents Claude Code loads into context (CLAUDE.md, MEMORY.md, skills, agents, commands, plugins) for outdated claims, contradictions, risky/ambiguous wording, and session-only leakage. Produces a prioritized report at .drift-reports/.
zforge
Durable state, contract and evidence layer for multi-session feature work: contract-bounded planning, autonomous phase orchestration with evidence-based acceptance, and ledger-versus-implementation review.
esp32-cpp
Evidence-first ESP32 firmware guidance for hardware, FreeRTOS, memory, peripherals, and field failures.
self-improving-agent
Curate auto-memory, promote learnings to CLAUDE.md and rules, extract proven patterns into reusable skills.
yarbrain
A reviewable Markdown second brain for Codex and Claude Code.