Best Claude skills for anthropic.
355 claude skills in the catalog work with anthropic. Each one links to its own page with its tools, install details and safety review.
Every one below is listed in ahel’s catalog and can be served to your AI today. Browse all Skills.
25 results
- View details
autonomous-agent-harnessDocs & knowledge
Lets your agent run tasks on a schedule, remember context across sessions, and work autonomously without you prompting each step.
Ready to connect★ 265k
- View details
cost-aware-llm-pipelineProductivity
Lets your agent cut LLM API costs by routing tasks to cheaper models, tracking spend against a budget, and caching prompts.
Ready to connect★ 265k
- View details
council-multi-modelAI & models
Lets your agent add an external Codex critique to a council claude skill decision draft before you make the final call.
Ready to connect★ 265k
- View details
ecc-tools-cost-auditAI & models
ECC claude skill that audits runaway PR creation, quota bypass, and premium-model cost leaks in the ECC Tools repo.
Ready to connect★ 265k
- View details
exa-searchSearch
Lets your agent search the web, find code examples, and research companies using the Exa search skill.
Ready to connect★ 265k
- View details
frontend-design-directionWeb & browsing
Guides your agent to apply a frontend claude skill mindset so web UIs it builds look polished and purposeful, not generic.
Ready to connect★ 265k
- View details
gan-style-harnessMedia
Lets your agent build whole applications autonomously by having one agent write the code and another strictly critique it until quality standards pass.
Ready to connect★ 265k
- View details
Algorithmic ArtAI & models
Lets your agent create algorithmic art skill projects — generative p5.js sketches with seeded randomness and interactive parameters.
Ready to connect★ 178k
- View details
brand-guidelinesMedia
Lets your agent style artifacts with Anthropic's brand guidelines skill, applying official colors and fonts to documents.
Ready to connect★ 178k
- View details
frontend-designMedia
Lets your agent apply a frontend design skill when building UI, with deliberate typography, color, and layout choices.
Ready to connect★ 178k
- View details
skill-creatorAI & models
Lets your agent create new skills, edit existing ones, and test how well they work.
Ready to connect★ 178k
- View details
ponytail-helpDev tools
Shows a quick-reference card of all ponytail claude skill modes and commands, so you can see what's available.
Ready to connect★ 144k
- View details
open-code-reviewDev tools
Open Claude skill lets your agent review Git changes and pull requests for bugs, security and performance issues.
Ready to connect★ 40k
- View details
creating-online-evaluationsMonitoring & ops
Guides your agent through creating Claude skill style online evaluations that automatically score AI responses in PostHog.
Ready to connect★ 40k
- View details
debugging-local-task-agent-runsCloud & infra
Guides debugging claude skill runs of local PostHog tasks by finding the sandbox and reading its wizard and agent logs.
Ready to connect★ 40k
- View details
debugging-mcp-analyticsDatabases & data
Lets your agent debug PostHog MCP Analytics with a debugging claude skill for missing events or wrong numbers.
Ready to connect★ 40k
- View details
debugging-signals-pipelineDatabases & data
Debugging claude skill for the PostHog signals pipeline: trace workflows, read sandbox logs, and diagnose failures.
Ready to connect★ 40k
- View details
exploring-llm-tracesFiles & storage
Lets your agent inspect and debug AI agent traces in PostHog, from a pasted trace or session URL.
Ready to connect★ 40k
- View details
finding-llm-gateway-migration-candidatesCloud & infra
Lets your agent find and rank code callers that could migrate from one LLM gateway to another, without changing anything.
Ready to connect★ 40k
- View details
implementing-warehouse-sourcesDatabases & data
Lets your agent build and extend PostHog data warehouse import sources that pull data from external APIs.
Ready to connect★ 40k
- View details
migrating-llm-gateway-callersAI & models
Walks your agent through moving a service's LLM calls from the old Python gateway to the shared Go AI gateway safely.
Ready to connect★ 40k
- View details
optimizing-clickhouse-and-hogql-queriesDatabases & data
Lets your agent diagnose and speed up slow ClickHouse and HogQL queries in PostHog.
Ready to connect★ 40k
- View details
setting-up-devboxDev tools
Lets your agent set up, start, connect to, and manage a remote PostHog devbox for development work.
Ready to connect★ 40k
- View details
writing-evalsCloud & infra
Lets your agent write and run eval suites for PostHog's AI harness, like writing claude skill tests.
Ready to connect★ 40k
- View details
embedding-strategiesSearch
Guides your agent in choosing and tuning embedding models for semantic search.
Ready to connect★ 40k
