Results for “agent-evaluation”

Evaluates execution transcripts and output files against a list of expectations, assigning pass/fail verdicts with cited evidence and critiquing the assertions themselves.

Design structured evaluation rubrics for scoring LLM and agent outputs — defining quality dimensions, scoring scales, hard gates, score descriptions, and edge cases. Load when the user asks to create an eval rubric, define evaluation criteria, design scoring dimensions, write an eval spec, or says "what should I evaluate", "design a rubric", "create eval criteria", "define quality dimensions", "evaluation rubric for", "how do I measure quality of". Sub-skill of eval-output orchestrator.

Build AI applications using the Azure AI Projects SDK for TypeScript, managing agents, connections, deployments, datasets, indexes, and evaluations.

Patterns and techniques for evaluating and improving AI agent outputs. Use this skill when: - Implementing self-critique and reflection loops - Building evaluator-optimizer pipelines for quality-critical generation - Creating test-driven code refinement workflows - Designing rubric-based or LLM-as-judge evaluation systems - Adding iterative improvement to agent outputs (code, reports, analysis) - Measuring and improving agent response quality

Optimizes long-running agent sessions with structured context compression, summarization, and durable handoff summaries that preserve decisions, files, risks, and next actions.

Autonomous experiment loop that optimizes any file by a measurable metric. Inspired by Karpathy's autoresearch. The agent edits a target file, runs a fixed evaluation, keeps improvements (git commit), discards failures (git reset), and loops indefinitely. Use when: user wants to optimize code speed, reduce bundle/image size, improve test pass rate, optimize prompts, improve content quality (headlines, copy, CTR), or run any measurable improvement loop. Requires: a target file, an evaluation command that outputs a metric, and a git repo.

Autonomous experiment loop that optimizes any file by a measurable metric. Inspired by Karpathy's autoresearch. The agent edits a target file, runs a fixed evaluation, keeps improvements (git commit), discards failures (git reset), and loops indefinitely. Use when: user wants to optimize code speed, reduce bundle/image size, improve test pass rate, optimize prompts, improve content quality (headlines, copy, CTR), or run any measurable improvement loop. Requires: a target file, an evaluation command that outputs a metric, and a git repo.

Validate process decomposition and architecture design quality before execution begins. Load when the setup-evaluator agent fires (automatic for agent-chain tasks), or when user says "evaluate this setup", "check the decomposition", "validate the architecture", "is this plan sound", "review the agent design". Catches structural errors, missing knowledge, unrealistic step ordering, and topology mismatches. Does NOT modify — only evaluates.

Use to assemble a fair, evidence-backed coaching pack for a support agent's one-to-one from QA evaluations and conversation history. Trigger for "prepare a coaching session for X", "what areas does X need to improve", "areas of markdown for this agent", "what coaching opportunities stand out", "build a coaching agenda from these tickets", or preparing a weekly or monthly agent review.

Evaluate everything the PA agent manages — tasks, skills, PA network health, billing, calendar connections, and memory quality. Use when: owner asks for an evaluation, wants to know what's working and what isn't, or requests a performance report. Combines supervisor status with quality scoring.

Provides a formal evaluation framework for Claude Code sessions, implementing eval-driven development (EDD) principles to define pass/fail criteria, measure reliability with pass@k metrics, and create regression test suites.

Instrument a shipped product's AI agents with tracing and observability so you can see what they did, why outputs happened, and what each run cost. Plain-language primer plus free-tier-first backend selection (Langfuse, Phoenix, LangSmith, Braintrust) and OpenTelemetry/OpenInference instrumentation. Load when the user asks to add observability, add tracing, instrument my agents, see what my agent is doing in production, set up Langfuse or Phoenix or LangSmith, debug why my agent gave a bad answer, or track LLM cost per request. Also fires when agent-system-architecture or setup-evaluation requires an observability plan for an agent-chain product. NOT for tracing the coding agent itself — that is run-trace. Precondition for runtime-learning-loop.

Orchestrator for the eval-output skill suite — evaluate LLM and agent outputs for quality, accuracy, helpfulness, and safety using structured rubrics and LLM-as-judge techniques. Load when the user says "evaluate this output", "score this response", "run an eval", "LLM as judge", "evaluate agent output", "how good is this response", "rate this answer", "eval this", or provides an LLM output that should be assessed for quality. Single entry point for all output evaluation workflows.

Build and improve AI or probabilistic software through evaluation-driven development. Use for LLM applications, agents, prompts, RAG, tool use, classifiers, model migrations, quality regressions, golden datasets, LLM-as-judge rubrics, benchmarks, or requests to add evals and measurable release gates. Pair with TDD for deterministic code; do not use as the primary guide for ordinary unit testing without model behavior.

Guardian Angel gives AI agents a moral conscience rooted in Thomistic virtue ethics. Rather than relying solely on rule lists, it cultivates stable virtuous dispositions— prudence, justice, fortitude, temperance—that guide every interaction. The foundation is caritas: willing the good of the person you serve. From this flow the cardinal virtues as practical habits of right action and sound judgment. v3.0 introduced virtue-based disposition as the primary evaluation layer, providing deeper coherence than checklists alone. The agent's character becomes the safeguard. v3.1 adds: Plugin enforcement layer with before_tool_call hooks, approval workflows for ambiguous cases, and protections for sensitive infrastructure actions.

Route Firebase AI feature work into either direct app/client Firebase AI Logic SDK integration or a server-owned Genkit workflow. Use when a web, mobile, backend, or full-stack feature needs model calls, typed outputs, reusable flows, tools, retrieval, prompt files, evals, observability, or deployment. Choose client-ai-logic, flow-foundation, tool-and-agent, retrieval-and-prompt, evaluation-and-observability, deployment-runtime, or comparison-or-fallback; route Firebase platform/operator work to `firebase-cli` and broad framework comparisons to `survey`.

AI/LLM red-team assessment using the OWASP LLM Top 10 (2025) + OWASP AI Testing Guide (AITG v1, Nov 2025) frameworks, plus OWASP MCP Top 10 runtime testing for agentic/MCP targets. Tests prompt injection, jailbreaks, system prompt leakage, sensitive data extraction, excessive agency, improper output handling, model extraction, content bias, evasion, membership inference, MCP token exposure, MCP command injection, and more. Uses four tools in combination: FuzzyAI (single-turn jailbreak fuzzing), PyRIT (multi-turn orchestrated attacks), Garak (probe-based vulnerability scanning), and promptfoo (plugin-based red-team evaluation). Each tool covers different OWASP categories; running them together gives systematic coverage. Includes a conditional MCP reconnaissance phase and a post-access AI infrastructure phase (chained from /post-exploit). Produces: OWASP LLM Top 10 + AITG + MCP coverage matrix, findings per category, architecture diagram of the AI system, PoCs for confirmed exploits. Chains into /gh-export for