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 benyki/gtm-engine --skill engine-loopgit clone --depth 1 https://github.com/benyki/gtm-engineWrote 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/benyki/gtm-engine/engine-loop)<a href="https://agentmods.dev/skills/benyki/gtm-engine/engine-loop"><img src="https://agentmods.dev/badge/skills/benyki/gtm-engine/engine-loop/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/benyki/gtm-engine/engine-loop"><img src="https://agentmods.dev/badge/skills/benyki/gtm-engine/engine-loop.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 | $0.00118 | $0.04615 |
| Opus 5.5 | $0.00047 | $0.01846 |
| Sonnet 5.5 | $0.00024 | $0.00923 |
| Haiku 4.5 | $0.00012 | $0.00462 |
Grade A, and why
engine-loop 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 2mo 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 — 231 lines — stays where its author put it; the contents beside it link to each section on GitHub.
engine-loop
Not an engine. Seo, social, video, outreach (and any custom engine) are
the engines. engine-loop is the framework underneath them: it turns their
runs into learning so those engines grow and compound over time.
Four jobs; run them in this order, because each depends on the one before.
| Job | Cadence | Command |
|---|---|---|
| fetch metrics | daily | due_metrics.py → read → runlog.py metric |
| score + challenge | weekly | score_arms.py, then act on the verdict |
| generate inputs | weekly | see What to make next |
| report | weekly | render_report.py |
Call them by full path — python3 ~/.agents/skills/engine-loop/scripts/<script>.py — from anywhere. That's the canonical store every install writes to, so the path is the same on every machine. Two shorter forms look tempting and both fail: scripts/… only resolves if the cwd is the skill folder, which is the one place the home isn't found from, and skills/… needs the home symlink, which is optional. The scripts locate the home from the current directory, so run them from the home or anywhere inside it — or pass --home <path> from outside. weekly.sh chains the deterministic ones and resolves its own paths.
The home is one folder per engine, plus shared/. Each engine folder is self-contained — its own engine.json (type, goal, primary metric), experiments.json, sources.json, templates/, inputs/, runs/, reports/. The loop scripts operate per folder and default to all of them; --engine <folder> scopes to one. Nothing is pooled across engines, so an agent rewriting one engine cannot break another — and two engines of the same type (outreach/ and outreach-investors/) are just two folders.
Starting fresh in a home you don't know? Read each engine's reports/latest.json first (and shared/insights.md for what the engines have learned from each other). It gives you the last period's runs, what shipped, the metric totals and their sources, every live experiment with its verdict, and how many runs are still owed a number — as data. Don't reconstruct that by reading CSVs.
Paths in this file: shared/… means the gtm home (~/gtm by default, or
$GTM_HOME); templates/, inputs/, runs/ and reports/ mean the engine
folder you're running, wherever it lives. The scripts resolve both through
~/gtm/engines.json, so read them as names rather than literal paths.
What ships with it
11 files 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.
- references/ab-testing.md 15 KB
- references/advanced.md 2.9 KB
- references/fetching-data.md 8.9 KB
- references/scheduling.md 6.1 KB
- scripts/assign_arm.py 9.9 KB runs code
- scripts/due_metrics.py 6.2 KB runs code
- scripts/gtmfind.py 10 KB runs code
- scripts/render_report.py 12 KB runs code
- scripts/runlog.py 10 KB runs code
- scripts/score_arms.py 9.6 KB runs code
- scripts/weekly.sh 1.6 KB runs code
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.
- 2mo ago First seen · 231 lines · 118 tokens per session scan A b85b4925131d
engine-loop is a skill published in the GitHub repository benyki/gtm-engine (4 stars, last pushed 6d ago), licensed MIT. It adds 118 tokens to every session and 4,615 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 2026-08-31.
engine-loop compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| engine-loopbenyki/gtm-engine | Skill | 4 | 6d ago | MIT | A | 118 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 134k | today | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 99k | today | Apache-2.0 | A | 19 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | today | MIT | A | 200 | Claude Code, Codex |
| systematic-debuggingobra/superpowers | Skill | 314k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 103 | Claude Code |