Coding standards, frameworks, refactoring, and code review (16.6K skills)
Define shared project vocabulary before coding. Gather terms from requests, code, and conversations, distinguish meanings across system areas, and create a CONTEXT.md glossary plus records explaining modeling choices that need justification.
Clean up unnecessary AI-generated additions in a branch diff, including obvious comments, excessive abstractions, and redundant defensive code. Preserve behavior through focused edits and verification, or review SKILL.md files for outdated and repetitive guidance.
Review Postgres queries, schema design, and access controls with prioritized guidelines. Use SQL examples and diagnostic guidance to investigate slow queries, manage connections, plan database changes, and check row-level security rules.
Implement and troubleshoot Supabase databases, authentication, and storage using current documentation. The skill guides access control checks, schema changes, log inspection, and test queries to verify that fixes work.
Write commit messages using Conventional Commits, a format that labels the kind of change and its scope. Produce a concise subject and, when needed, a body explaining what changed and why.
Review Python code against a shared checklist covering code quality, input validation, and tests. Get pass or fail findings for each area, plus a summary of items that need attention before merging.
Delegate clearly defined tasks, such as typo fixes, tests, and simple refactoring, through CCS CLI. The skill selects a configured model profile, adds relevant file context and success criteria, runs the task, and reports execution details.
Diagnose slow Attio requests and growing sync queues by measuring query complexity, pagination, concurrency, retries, and cache behavior. Compare a targeted change against a representative baseline while checking data correctness, freshness, and rollback thresholds.
Analyze Attio integrations that hit request limits and design separate read and write budgets, controlled queues, and bounded retries. Produce a request control policy with load test evidence and clear rollback thresholds.
Triage failures across AssemblyAI REST requests, transcripts, Streaming v3, webhooks, and LLM Gateway. Separate transport, authentication, quota, processing, and callback issues, then report a retry, input fix, capacity action, or escalation using content-free evidence.
Use this skill to plan and validate migrations from AssemblyAI Streaming v2 and LeMUR to Streaming v3 and LLM Gateway. It guides endpoint and message changes, output comparisons, synthetic test cases, canary rollout, rollback checks, and legacy cleanup.
Set up REST authentication for an Attio integration by choosing a workspace API key or OAuth for multiple workspaces. Map required scopes to endpoints, protect and rotate tokens, and verify access with a read-only request.
Build authenticated handlers for AssemblyAI transcription callbacks, notifications sent to an application. Validate incoming events, prevent duplicate actions, acknowledge delivery promptly, and retrieve transcript results separately while keeping sensitive content out of validation reports.
Create or revise reusable agent capabilities such as Agent Skills, plugins, subagents, MCP servers, hooks, and marketplace entries. Choose the correct format for each host, apply least privilege, and keep portable instructions separate from host-specific packaging.
Set up a repeatable local workflow for testing Attio data mappings and API responses. Use sanitized test data and offline checks by default, with explicit controls for live writes and records of test resource cleanup.
Use this skill to migrate an existing Attio integration when endpoints, schemas, client versions, or architecture change. It guides a contract comparison, compatibility work, shadow checks, a limited canary rollout, reconciliation, and a tested rollback plan.
Design a production Attio integration from existing repository code. Plan workspace-specific credentials, queued writes, verified event notifications, duplicate handling, and data reconciliation, then document component ownership, rollout steps, and rollback procedures.
Prepare a small, redacted evidence bundle for an Attio integration failure. It captures request details, relevant schema context, retry behavior, and environment metadata while screening out CRM records, personal data, and secrets.
Configure AssemblyAI authentication and regional endpoints for an integration. Keep project keys on trusted servers, design temporary streaming tokens for browser or mobile clients, and document configuration checks without exposing credentials or audio content.
Design a typed adapter around the official AssemblyAI SDK so application code stays insulated from vendor changes. Define stable request and result types, regional routing, retries, polling, and streaming lifecycles while keeping credentials and sensitive audio or transcript content out of evidence.
Plan AssemblyAI throughput and handle request limit errors by reviewing account capacity, separating operation queues, and bounding retries. Use callbacks, delayed retries, and transcript ID checks to control polling load and repeated submissions.
Build or refactor a typed client for Attio’s REST API, with workspace-aware credentials, endpoint-specific pagination, structured errors, and bounded retries. The skill helps define tested integration boundaries and keeps the REST client distinct from Attio’s in-product App SDK.
Check a new Attio token, workspace, or integration path with one read-only REST request. The skill verifies authentication and the response format by listing objects, while keeping write scopes out of the proof and secrets out of logs.
Diagnose Attio REST API failures using response status, structured error details, and endpoint documentation. Identify authentication, permission, validation, and rate-limit problems, then apply a minimal correction and verify it while keeping tokens and customer data out of diagnostic records.