This repository contains skills to use with Claude Code.
Claude Skills are customizable workflows that teach Claude how to perform specific tasks according to your unique requirements. Skills enable Claude to execute tasks in a repeatable, standardized manner across all Claude platforms.
The goal of Synalinks skills is to teach Claude to use the Synalinks framework correctly. Synalinks is Keras-inspired, so without guidance LMs tend to mix Keras / LangChain / DSPy syntax — producing plausible-looking but broken code. These skills constrain Claude to idiomatic Synalinks usage.
Each skill targets one slice of the framework. Claude auto-activates the relevant skill(s) based on what you're asking about.
| Skill | When it activates |
|---|---|
| synalinks-core | DataModel, Field, Input, JSON operators (+ & | ^ ~), synalinks.ops, configuration, LanguageModel/EmbeddingModel basics |
| synalinks-programs | Program class, four building APIs (Functional / Sequential / Subclassing / Mixed), multi-input/output graphs, build/call lifecycle, save/load, summary, get_module, plot_program, custom serialization |
| synalinks-modules | Generator, ChainOfThought, SelfCritique, Identity, PythonSynthesis, SequentialPlanSynthesis, custom modules via subclassing |
| synalinks-control-flow | Decision, Branch, parallel branches, self-consistency, XOR input/output guards, And/Or modules, branch merging |
| synalinks-agents | FunctionCallingAgent, ToolCalling, Tool definitions, MCP integration (MultiServerMCPClient), trajectories |
| synalinks-knowledge | KnowledgeBase (DuckDB), EmbedKnowledge, UpdateKnowledge, RetrieveKnowledge, RAG/KAG, hybrid search, Entity/Relation graphs |
| synalinks-training | compile() / fit() / evaluate() / predict(), callbacks, ProgramCheckpoint, training workflow |
| synalinks-rewards | ExactMatch, CosineSimilarity, LMAsJudge, ProgramAsJudge, MeanRewardWrapper, F1Score, custom rewards/metrics, masking |
| synalinks-optimizers | RandomFewShot, OMEGA, Dominated Novelty Search, mutation/crossover, quality-diversity tuning |
| synalinks-providers | Provider prefixes (openai, anthropic, groq, openrouter, cohere, deepseek, together_ai, bedrock, doubleword, hosted_vllm, ...), local servers (LMStudio/vLLM), OpenRouter embeddings |
| synalinks-datasets | Built-in datasets (gsm8k, hotpotqa, arcagi), custom iterable datasets, visualization (plot_program, plot_history, plot_metrics_*) |
Each skill folder contains:
SKILL.md— frontmatter + scannable usage guidereferences/— deep-dive reference docsscripts/— runnable example scripts
Clone the repository — every install path below copies, zips, or symlinks the
synalinks-*/ folders out of this checkout, so keep it somewhere stable.
git clone https://github.com/SynaLinks/synalinks-skills.git
cd synalinks-skillsPick one of the install targets below depending on where you use Claude.
Skills live in one of two locations:
| Scope | Path | When to use |
|---|---|---|
| User (all projects) | ~/.claude/skills/ |
You always want the skills available, regardless of which repo you're in. |
| Project (one repo) | <repo>/.claude/skills/ |
Limit the skills to a single Synalinks project; commit them with the repo so teammates pick them up automatically. |
Install at the user level (recommended for solo use):
mkdir -p ~/.claude/skills/
cp -r synalinks-* ~/.claude/skills/Or install for a single project:
mkdir -p /path/to/your/synalinks-project/.claude/skills/
cp -r synalinks-* /path/to/your/synalinks-project/.claude/skills/Tip: prefer
ln -s "$PWD"/synalinks-* ~/.claude/skills/if you want updates fromgit pullto flow through automatically.
Verify the install — start Claude Code and run the /skills slash command;
the 11 synalinks-* skills should appear in the list. You can also peek at a
skill's frontmatter directly:
head ~/.claude/skills/synalinks-core/SKILL.mdClaude auto-activates the relevant skill(s) for your task based on each
SKILL.md's description field — no manual selection needed.
These uploads expect a single .skill archive (a zip of one skill folder
with SKILL.md at the root). Build one archive per skill:
for d in synalinks-*/; do zip -r "${d%/}.skill" "$d"; doneThen in the Claude interface, click the skill icon (🧩), choose "Upload
skill", and upload each .skill file. Repeat for every skill you want
available. See Using skills with Claude
for the latest UI walkthrough.
cd synalinks-skills
git pull
cp -r synalinks-* ~/.claude/skills/ # if you copied (skip if you symlinked)For Claude.ai / Desktop, regenerate the .skill archives and re-upload the
ones that changed. Restart Claude Code (or run /skills → reload) so the
new metadata is picked up.
rm -rf ~/.claude/skills/synalinks-*For Claude.ai / Desktop, remove each skill from the skill icon (🧩) menu.
These skills are licensed under Apache 2.0, like Synalinks framework.
See the LICENSE file for full details.
Contributions are welcome! Please feel free to submit a Pull Request.
These skills have been created by Ramiro Salas the CTO of Hexatropian an active early member of Synalinks community.
- Synalinks Framework - The neuro-symbolic AI framework these skills are designed for