Skill UI

Automates the CI/CD process for Glean enterprise search connectors. This tool runs comprehensive unit tests with mocked responses on every Pull Request, and executes integration tests against a staging Glean instance upon merging to the main branch. It validates connector data transforms, indexing API calls, and performs search relevance regression testing to ensure high search quality.

This harness provides comprehensive, reproducible evaluation pipelines for complex LLM chains and agents (LangChain/LangGraph 1.0). It integrates golden dataset management, LangSmith evaluation, RAGAS metrics, deepeval LLM-as-judge, and structured agent trajectory analysis. Use it when establishing quality benchmarks for new chains, diagnosing performance regressions after model switches, or implementing CI/CD gates to prevent quality drops.

A systematic, multi-phase guide for tackling hard bugs and performance regressions. It mandates building a robust feedback loop (e.g., failing tests, harnesses), formulating falsifiable hypotheses, targeted instrumentation, and rigorous regression testing. This methodology ensures that debugging efforts are structured, efficient, and highly reproducible, making it ideal for complex software defects.

This skill monitors SEO changes over time by capturing baselines of critical elements like title tags, meta descriptions, and Core Web Vitals. It detects regressions and tracks history, acting as Git for SEO. Use it to check for deployment breaks, compare before and after states, or monitor SEO stability. Commands include baseline, compare, and history.

Diagnoses full-stack failure modes in agent and LLM applications, including wrapper regressions, memory contamination, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings and code-first fixes for production agent systems.

This tool empowers users to query, filter, and aggregate logs from Datadog directly from the command line using the Composio CLI. It bypasses the graphical UI, providing structured JSON output for complex analysis, cross-service correlation, and automated reporting. Ideal for deep dive investigations, tracking latency regressions, or building scheduled operational digests.

Perform a priori power analysis to determine required sample sizes for studies. Supports closed-form calculations for t-tests, ANOVA, and regression, plus Monte Carlo simulations for complex models like mixed models and survival analysis. Helps justify sample sizes for grants and protocols by calculating minimum detectable effects and power curves.

This skill enables the creation of a robust 'harness' system for coding agents within a repository. It addresses the need to prevent repeated AI-generated coding mistakes by implementing durable instructions, guardrails, regression tests, and failure memory. It treats the target repository as the source of truth, ensuring consistency and maintaining high reliability across all automated contributions.

This skill outlines a systematic, multi-phase workflow for diagnosing hard bugs and performance regressions in complex codebases. It emphasizes the critical necessity of building a tight, deterministic feedback loop (Phase 1) before moving to hypothesis generation. The process guides users through reproducing the failure, minimizing the scope of the issue, and forming actionable, ranked hypotheses.

An automated framework for continuous, repeatable red-teaming of LLM applications within CI/CD pipelines. It integrates Promptfoo and DeepTeam to test against recognized vulnerabilities like OWASP LLM Top 10, OWASP Agentic, and MITRE ATLAS (e.g., Prompt Injection, Jailbreaks). This ensures that security vulnerabilities do not regress after code or model changes, providing a critical merge-blocking security gate for production LLM apps.

A systematic, non-mutating investigation designed for complex software failures, including bugs, regressions, crashes, and flaky behavior. This skill launches four specialized, read-only sub-agents in parallel to perform deep root-cause analysis, trace code paths, pinpoint the smallest failing seam, and predict the quickest path to confirming or fixing the issue. It is designed purely for diagnosis and hypothesis generation.

Utilize this skill to assess the health, adoption, and performance metrics of published EAS Updates directly via the CLI. It reports critical data points including crash rates, unique user counts, install volumes, payload size, and the split between embedded and OTA users, making it essential for post-publish monitoring and regression detection.