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 thinkwee/AgentsMeetRL --skill agents-meet-rlgit clone --depth 1 https://github.com/thinkwee/AgentsMeetRLWrote 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/thinkwee/agentsmeetrl/agents-meet-rl)<a href="https://agentmods.dev/skills/thinkwee/agentsmeetrl/agents-meet-rl"><img src="https://agentmods.dev/badge/skills/thinkwee/agentsmeetrl/agents-meet-rl/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/thinkwee/agentsmeetrl/agents-meet-rl"><img src="https://agentmods.dev/badge/skills/thinkwee/agentsmeetrl/agents-meet-rl.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00220 | $0.00916 |
| Opus 5.5 | $0.00088 | $0.00366 |
| Sonnet 5.5 | $0.00044 | $0.00183 |
| Haiku 4.5 | $0.00022 | $0.00092 |
Grade A, and why
agents-meet-rl 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 11d 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.
The source is not reproduced here
No licence file
A repository with no LICENSE is all rights reserved by default, so the body is not copied here. The metadata, the measurements and the link are.
What ships with it
60 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.
- database.json 316 KB
- evals/evals.json 3.3 KB
- problems/_INDEX.md 13 KB
- problems/evaluation/benchmark-pitfalls.md 4.5 KB
- problems/evaluation/canary-eval.md 8.3 KB
- problems/evaluation/data-contamination.md 3.4 KB
- problems/evaluation/inference-time-scaling.md 2.7 KB
- problems/evaluation/llm-judge-evaluation.md 2.6 KB
- problems/evaluation/long-horizon-eval.md 3.6 KB
- problems/evaluation/multi-agent-eval.md 2.9 KB
- problems/evaluation/ood-evaluation.md 2.9 KB
- problems/evaluation/pass-at-k.md 2.9 KB
- problems/evaluation/reproducibility.md 4.1 KB
- problems/evaluation/statistical-significance.md 3.0 KB
- problems/evaluation/train-eval-mismatch.md 3.4 KB
- problems/research-workflow/ablation-design.md 3.8 KB
- problems/research-workflow/baseline-selection.md 4.6 KB
- problems/research-workflow/compute-budget.md 2.5 KB
- problems/research-workflow/contribution-framing.md 4.7 KB
- problems/research-workflow/data-curation.md 5.2 KB
- problems/research-workflow/domain-application.md 4.1 KB
- problems/research-workflow/hyperparam-search.md 8.5 KB
- problems/research-workflow/online-vs-offline-rl.md 8.5 KB
- problems/research-workflow/reproducing-papers.md 3.6 KB
- problems/training/algorithm-choice.md 15 KB
- problems/training/async-rollout-staleness.md 4.5 KB
- problems/training/catastrophic-forgetting.md 7.8 KB
- problems/training/code-swe-specific.md 5.2 KB
- problems/training/conversation-history-truncation.md 9.6 KB
- problems/training/credit-assignment-long-horizon.md 6.9 KB
- problems/training/discount-factor.md 2.1 KB
- problems/training/distributed-failures.md 8.9 KB
- problems/training/entropy-collapse.md 6.4 KB
- problems/training/env-flakiness.md 3.9 KB
- problems/training/format-reward-gaming.md 2.8 KB
- problems/training/grpo-knobs.md 4.8 KB
- problems/training/gui-specific.md 5.9 KB
- problems/training/importance-ratio-blowup.md 2.3 KB
- problems/training/kl-explosion.md 4.7 KB
- problems/training/kl-penalty-tuning.md 2.8 KB
- problems/training/length-blowup.md 4.8 KB
- problems/training/length-collapse.md 2.9 KB
- problems/training/llm-judge-pitfalls.md 5.4 KB
- problems/training/lora-rl.md 2.1 KB
- problems/training/memory-context-overflow.md 5.8 KB
- problems/training/mixed-precision.md 8.7 KB
- problems/training/multi-agent-credit-collapse.md 4.8 KB
- problems/training/nan-grads.md 2.6 KB
- problems/training/no-eos.md 3.0 KB
- problems/training/observation-truncation.md 3.4 KB
- problems/training/oom-during-training.md 2.5 KB
- problems/training/ppo-knobs.md 4.0 KB
- problems/training/prm-training.md 4.2 KB
- problems/training/reasoning-mode-rl.md 12 KB
- problems/training/retokenization-drift.md 4.9 KB
- problems/training/reward-mixing.md 5.7 KB
- problems/training/reward-not-increasing.md 4.9 KB
- problems/training/reward-up-eval-flat.md 6.8 KB
- problems/training/rl-from-base.md 11 KB
- problems/training/rollout-throughput.md 3.5 KB
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.
- 11d ago Changed baf42d1c39f8
- 1mo ago First seen · 67 lines · 220 tokens per session scan A 2ab5f45f4275
agents-meet-rl is a skill published in the GitHub repository thinkwee/AgentsMeetRL (1,864 stars, last pushed yesterday), with no licence file. It adds 220 tokens to every session and 916 once invoked, about $0.0009 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 2026-08-30.
Other skills, from other repositories
data-analyzer
Guidance for analyzing structured data, generating statistics and producing data-driven insights. Use when the user asks to analyze data, compute statistics, find patterns, or generate analytical reports.
yait-aichain
Build AI features in Python — an LLM call, a multi-step pipeline, the same work over many items, or an agent that uses tools (web search, page fetching, your own functions, any REST API or MCP server, a working folder), asks a human before risky actions, stays within a money budget, or pauses and resumes — on any…
ddd-tactical-design
Use when implementing domain logic, designing aggregates, entities, value objects, repositories, domain services, domain model design, ensuring data consistency, or when user mentions 'aggregate design', 'aggregate root', 'entity vs value object', 'domain event', 'repository pattern', 'domain service', 'data…
prompt-enhancer
MUST use before task execution. Clarify vague requirements, explore solutions, and ensure full understanding.
yait-aichain
Build AI features in Python — an LLM call, a multi-step pipeline, the same work over many items, or an agent that uses tools (web search, page fetching, your own functions, any REST API or MCP server, a working folder), asks a human before risky actions, stays within a money budget, or pauses and resumes — on any…
agent-v3-memory-specialist
Agent skill for v3-memory-specialist - invoke with $agent-v3-memory-specialist.