prompting-claude-models
Skill10 starsChanged 57 days agoNo licenceView on GitHub
Design prompts, system prompts, and agent scaffolding for LLM-backed features built on Claude models — LLM-as-judge and evaluation harnesses, RAG over literature or documents, extraction/classification pipelines, and multi-step or subagent agents. Use whenever the user is writing or debugging a prompt for a Claude model: "improve this system prompt", "my LLM judge is inconsistent", "Claude ignores my instructions / over-triggers a tool / is too verbose", "how do I structure this prompt", "why won't Claude use my tool", "tune effort/thinking for this agent", or turning a research workflow into an LLM call. This is prompt *design* craft; for API/SDK mechanics (model IDs, pricing, parameters, tool-use schemas, caching, token counting) use the claude-api reference instead, and for Python code quality use the python-* skills. Model-version-specific behavior changes often, so this skill grounds those claims in the live Anthropic docs rather than reciting them from memory.
In this repository89 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- apply-cold-email-drafterSkill
- apply-cv-builderSkill
- apply-cv-reviewerSkill
- apply-dossier-evaluatorSkill
- apply-interview-prepSkill
- apply-motivation-keeperSkill
- apply-package-auditorSkill
- apply-profile-readerSkill
- apply-program-fit-mapperSkill
- apply-program-theme-extractorSkill
- apply-recommendation-letter-strategistSkill
- apply-research-direction-mapperSkill
- apply-sop-writerSkill
- banking-ai-literature-mapperSkill
- banking-use-case-discoverySkill
- industry-problem-framingSkill
- pilot-and-evaluation-designSkill
- publication-grounded-solution-designSkill
- init-workspaceSkill
- concept-exercise-generatorSkill
- knowledge-debt-auditSkill
- naive-studentSkill
- professor-criticSkill
- professor-mentor-technical-teachingSkill
- whiteboard-peerSkill
- deck-beamer-academicSkill
- deck-beamer-proposal-reportSkill
- deck-business-reportSkill
- abstract-and-intro-writerSkill
- discussion-report-synthesizerSkill
- figure-table-plannerSkill
- method-section-writerSkill
- paper-argument-plannerSkill
- research-proposal-writerSkill
- research-talk-plannerSkill
- results-writeupSkill
- cross-domain-analogy-finderSkill
- gap-finderSkill
- gap-motivation-builderSkill
- paper-idea-and-scope-brainstormerSkill
- research-idea-stress-testSkill
- jax-training-recipeSkill
- paper-to-codeSkill
- pytorch-training-recipeSkill
- research-codebaseSkill
- citation-auditorSkill
- literature-triangulationSkill
- related-work-writerSkill
- survey-paper-writerSkill
- benchmark-and-baseline-selectorSkill
- experiment-designSkill
- hypothesis-and-ablation-plannerSkill
- model-eval-error-analysisSkill
- statistical-testing-guideSkill
- artifact-release-packagerSkill
- peer-review-writerSkill
- reproducibility-auditSkill
- reviewer-response-strategistSkill
- The first 60 shown.
The file
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.