2.7K skills · sorted by relevance
Fix bugs reported in GitHub issues through a reproduce-first process. The workflow guides code changes, build checks, verification, regression tests that check for recurring failures, and review before considering the fix complete.
Build quarterly product plans by translating company goals into product and team OKRs, meaning objectives with measurable results. Generate linked goal hierarchies, check alignment across teams, and review how responsibilities are distributed.
Test marketing prompts against shared cases, score required content, prohibited phrases, and output format, and compare alternatives. Keep version history and diffs, use reusable templates, and check that prompt or model updates preserve output quality.
Plan and analyze A/B tests, prepare tabular data, and train and evaluate prediction models. Estimate intervention effects from observational data, track model experiments with MLflow, and translate statistical findings into business decisions.
Design systems that collect, transform, and store data. Choose between scheduled and continuous processing, generate workflow configurations, validate data quality, and analyze performance using guidance for tools including Python, SQL, Spark, Airflow, dbt, and Kafka.
Prepare trained machine learning models for production with deployment, monitoring, and retraining workflows. Build RAG systems that retrieve documents for answers, and integrate language model APIs with retries, fallback providers, and cost tracking.
Review architecture and engineering growth plans through six questions covering capacity limits, technical debt, staffing, build-versus-buy costs, reliability, and security. Produce a structured verdict and three next steps before committing to major technical decisions.
Plan a Claude Managed Agent around one clear job, then route work through setup, launch preparation, evaluation, and scheduling. Produce execution plans and curl commands that users run with their own Anthropic API key.
Route customer success, sales, revenue operations, and contract requests to one of four specialist skills. Use it when a business request is unclear, so the agent selects the relevant workflow before starting the work.
Find repeated engineering tasks and turn them into agent workflows with clear checks, stopping rules, and approval boundaries. Compare published workflows, adapt suitable examples, or review existing ones for weak verification and uncontrolled repetition.
Explore features for a new product through product management, design, and engineering perspectives. Generate five ideas from each viewpoint, then select five priorities with reasoning and assumptions to test before developing the concept further.
Generate improvement ideas for an existing product from product management, design, and engineering perspectives. Produce five ideas per perspective, then select the top five based on goals, potential impact, feasibility, and differentiation, with reasons and assumptions to test.
Generate text and images through a reverse-engineered Gemini Web API, an interface reconstructed from the web service. Use reference images for descriptions or variations, and continue conversations across requests with saved sessions.
Document a solved and verified problem when its reasoning is missing from code, tests, and existing documentation. Save one durable lesson in the repository so future agents can find it and avoid repeating substantial investigation.
Create a manufacturing drawing PDF from a part’s 3D geometry, with projected views, measured dimensions, hole callouts, and a title block. Rerun the drawing script after model changes to update the sheet.
Audit an entire codebase for unnecessary complexity and dependencies. Receive a ranked list of code to delete, simplify, or replace with standard library and platform features, with file paths and estimated reductions.
Create a weekly engineering and product update from GitHub and Slack, combining shipped outcomes with contributor rankings, merged pull request counts, weekly changes, consecutive shipping weeks, and a thread for asynchronous team check-ins.
Use the P7 engineering workflow in Codex CLI through a command alias. It guides planning, implementation, and a three-question self-review, then requires command and output evidence before reporting completion.
Evaluate a machine learning GPU cluster during a cloud trial or acceptance test. Measure GPU computing performance, communication within and between machines, and local and shared storage, then compile the results into a dated Markdown report.
Helps design and evaluate prompts for more reliable AI responses. Covers clear instructions, input-output examples, task breakdown, and structured formats such as JSON, with guidance on comparing quality, cost, and response time across models.
Guides machine learning work from data preparation and model training to evaluation and production monitoring. Covers PyTorch, scikit-learn, experiment tracking, and reproducible workflows so teams can compare models and maintain them after deployment.
Write and review launch prompts for agents working autonomously for hours or days. Define successful deliverables, excluded partial results, stopping rules, worker handoffs, and evidence-based review requirements so completion can be assessed against explicit criteria.
Design the controls around autonomous agents: what they can edit, how results are evaluated, how progress is logged, and when human approval is required. Build research and experiment loops with clear recovery and review rules.
Design loops that improve an agent’s own prompts, workflows, or execution code from failure records. Define bounded edits, preserve candidate histories, and accept changes through independent tests that check improvements and regressions.