← All skills
post-training-slime
80 / 100 · Low risk
Fine to try. Some of what it can do is expected: a coding skill usually runs build and test commands.
By davila7/claude-code-templates (GitHub) · coding · 33,256 stars · scanned at commit a6e6f93
Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data generation workflows, or needing tight Megatron-LM integration for RL scaling.
Works with any agent that supports SKILL.md skills.
What this skill can do
Runs commands on your computer
The author does not mention this. Expected for a coding skill.
Found in 18 places.
Show 15 more
- cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 132
ls scripts/models/ - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 142
python train.py \ - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 181
python train_async.py \ - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 194
--async-buffer-size 4 # Number of rollouts to buffer - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 239
python train.py \ - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 258
--tensor-model-parallel-size 2 - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 266
--sglang-mem-fraction-static 0.8 - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 274
--actor-num-nodes 1 - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 350
--use-fault-tolerance - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 366
--update-weights-interval 5 - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 379
--recompute-activations - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 395
--num-data-workers 4 - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 424
python train.py \ - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 446
--custom-rm-path custom_rm.py - cli-tool/components/skills/ai-research/post-training-slime/SKILL.md, line 452
--eval-prompt-data aime /path/to/aime.jsonl \
Downloads and installs other software
The author does not mention this.
Found in 3 places.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'post-training-slime' ] || [ -L ~/.agents/skills/'post-training-slime' ]; then echo 'A skill folder named post-training-slime already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/davila7/claude-code-templates' "$d" && git -C "$d" checkout --quiet 'a6e6f93e9a4baaa6ff337aee5505bf5c2bae4ce6' -- 'cli-tool/components/skills/ai-research/post-training-slime' && mkdir -p ~/.agents/skills && cp -R "$d"/'cli-tool/components/skills/ai-research/post-training-slime' ~/.agents/skills/'post-training-slime' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'post-training-slime' ~/.claude/skills/'post-training-slime' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for davila7/claude-code-templates shows the score of its lowest-scoring skill.

[](https://zalavijaysinh.github.io/skilltrust/repos/davila7/claude-code-templates/)
Score history
| Date | Version | Score |
|---|
| 2026-11-10 | a6e6f93 | 80 |
| 2026-11-08 | b086f26 | 80 |
| 2026-11-07 | 395f30e | 80 |