model-evaluation-benchmark

Included with Lifetime

$97 forever

Automated reproduction of comprehensive model evaluation benchmarks following the Benchmark Suite V3. Auto-activates for model benchmarking, comparison evaluation, or performance testing between AI models.

General

What this skill does

# Model Evaluation Benchmark Skill

**Purpose**: Automated reproduction of comprehensive model evaluation benchmarks following the Benchmark Suite V3 reference implementation.

**Auto-activates when**: User requests model benchmarking, comparison evaluation, or performance testing between AI models in agentic workflows.

## Skill Description

This skill orchestrates end-to-end model evaluation benchmarks that measure:

- **Efficiency**: Duration, turns, cost, tool calls
- **Quality**: Code quality scores via reviewer agents
- **Workflow Adherence**: Subagent calls, skills used, workflow step compliance
- **Artifacts**: GitHub issues, PRs, documentation generated

The skill automates the entire benchmark workflow from execution through cleanup, following the v3 reference implementation.

## When to Use

✅ **Use when**:

- Comparing AI models (Opus vs Sonnet, etc.)
- Measuring workflow adherence
- Generating comprehensive benchmark reports
- Need reproducible benchmarking

❌ **Don't use when**:

- Simple code reviews (use `reviewer`)
- Performance profiling (use `optimizer`)
- Architecture decisions (use `architect`)

## Execution Instructions

When this skill is invoked, follow these steps:

### Phase 1: Setup

1. Read `tests/benchmarks/benchmark_suite_v3/BENCHMARK_TASKS.md`
2. Identify models to benchmark (default: Opus 4.5, Sonnet 4.5)
3. Create TodoWrite list with all phases

### Phase 2: Execute Benchmarks

For each task × model:

```bash
cd tests/benchmarks/benchmark_suite_v3
python run_benchmarks.py --model {opus|sonnet} --tasks 1,2,3,4
```

### Phase 3: Analyze Results

1. Read all result files: `~/.amplihack/.claude/runtime/benchmarks/suite_v3/*/result.json`
2. Launch parallel Task tool calls with `subagent_type="reviewer"` to:
- Analyze trace logs for tool/agent/skill usage
- Score code quality (1-5 scale)
3. Synthesize findings

### Phase 4: Generate Report

1. Create markdown report following `BENCHMARK_REPORT_V3.md` structure
2. Create GitHub issue with report
3. Archive artifacts to GitHub release
4. Update issue with release link

### Phase 5: Cleanup (MANDATORY)

1. Close all benchmark PRs: `gh pr close {numbers}`
2. Close all benchmark issues: `gh issue close {numbers}`
3. Remove worktrees: `git worktree remove worktrees/bench-*`
4. Verify cleanup complete

See `tests/benchmarks/benchmark_suite_v3/CLEANUP_PROCESS.md` for detailed cleanup instructions.

## Example Usage

```
User: "Run model evaluation benchmark"Assistant: I'll run the complete benchmark suite following the v3 reference implementation.

[Executes phases 1-5 above]

Final Report: See GitHub Issue #XXXX
Artifacts: https://github.com/.../releases/tag/benchmark-suite-v3-artifacts
```

## References

- **Reference Report**: `tests/benchmarks/benchmark_suite_v3/BENCHMARK_REPORT_V3.md`
- **Task Definitions**: `tests/benchmarks/benchmark_suite_v3/BENCHMARK_TASKS.md`
- **Cleanup Guide**: `tests/benchmarks/benchmark_suite_v3/CLEANUP_PROCESS.md`
- **Runner Script**: `tests/benchmarks/benchmark_suite_v3/run_benchmarks.py`

---

**Last Updated**: 2025-11-26
**Reference Implementation**: Benchmark Suite V3
**GitHub Issue Example**: #1698

Files: 1

Size: 3.4 KB

Complexity: 12/100

Category: General

Source: https://github.com/rysweet/amplihack/tree/main/amplifier-bundle/skills/model-evaluation-benchmark

Related in General

modeling-omnistudio-epc-catalog

Included

Salesforce Industries CME EPC product-modeling skill for Product2-based catalog creation. Use when creating EPC products, configuring product attributes, building offer bundles with Product Child Items, or reviewing EPC DataPack JSON metadata for product catalog changes. TRIGGER when: user creates or updates Product2 EPC records, AttributeAssignment payloads, AttributeMetadata/AttributeDefaultValues, Offer bundles, or ProductChildItem relationships. DO NOT TRIGGER when: designing OmniScripts/FlexCards/Integration Procedures (use building-omnistudio-omniscript, building-omnistudio-flexcard, or building-omnistudio-integration-procedure), implementing Apex business logic (use generating-apex), or troubleshooting deployment pipelines (use deploying-metadata).

Generalscripts

relationship-science-coach

Included

Use this skill for direct, practical adult relationship coaching: couples conflict, repair, trust, marriage, dating, flirting, attachment patterns, emotional connection, sex, desire differences, eroticism, kink negotiation, affection, love languages, breakups, and long-term passion. Draw on Gottman, EFT and Hold Me Tight, attachment science, modern sex research, Perel, Nagoski, Kerner, Schnarch, Love and Stosny, and flexible love-language tools. Be concrete and low-hedge. Redirect only for imminent danger, abuse, coercive control, minors, non-consent, self-harm, stalking, or medical/legal/psychiatric decisions.

Generalscripts

building-sf-integrations

Included

Salesforce integration architecture and runtime plumbing with 120-point scoring. Use this skill to set up Named Credentials, External Credentials, External Services, REST/SOAP callout patterns, Platform Events, and Change Data Capture. TRIGGER when: user sets up Named Credentials, External Services, REST/SOAP callouts, Platform Events, CDC, or touches .namedCredential-meta.xml files. DO NOT TRIGGER when: Connected App/OAuth config (use configuring-connected-apps), Apex-only logic (use generating-apex), or data import/export (use handling-sf-data).

Generalscripts

venue-templates

Included

Access comprehensive LaTeX templates, formatting requirements, and submission guidelines for major scientific publication venues (Nature, Science, PLOS, IEEE, ACM), academic conferences (NeurIPS, ICML, CVPR, CHI), research posters, and grant proposals (NSF, NIH, DOE, DARPA). This skill should be used when preparing manuscripts for journal submission, conference papers, research posters, or grant proposals and need venue-specific formatting requirements and templates.

Generalscripts

let-fate-decide

Included

Draws the 12 Houses of the Zodiac Tarot spread to inject entropy into planning when prompts are vague, ambiguous, or casually delegated. Interprets the spread to guide next steps. Use when the user says 'let fate decide', 'YOLO', 'whatever', 'idk', or other nonchalant phrases, makes Yu-Gi-Oh references, or when you are about to arbitrarily pick between multiple reasonable approaches. Prefer over ask-questions-if-underspecified when the user's tone is casual or playful rather than precision-seeking.

Generalscripts

net-ops

Included

Cross-platform network troubleshooting (Windows, macOS, Linux) via local or remote shell. Use for: DNS broken, can't resolve hostnames, nslookup/dig works but apps fail, NRPT, WFP, scutil, /etc/resolver, systemd-resolved, /etc/resolv.conf, NetworkManager, VPN DNS leak residue (ProtonVPN/Mullvad/WireGuard/AnyConnect), AV/firewall blocking DNS or DoH, Tailscale DNS interaction, intermittent connectivity, remote diagnostics over SSH.

Generalscripts