Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add hamzabellouch/agent-skills --skill agent-platform-eval-flywheelgit clone --depth 1 https://github.com/hamzabellouch/agent-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/hamzabellouch/agent-skills/agent-platform-eval-flywheel)<a href="https://agentmods.dev/skills/hamzabellouch/agent-skills/agent-platform-eval-flywheel"><img src="https://agentmods.dev/badge/skills/hamzabellouch/agent-skills/agent-platform-eval-flywheel/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/hamzabellouch/agent-skills/agent-platform-eval-flywheel"><img src="https://agentmods.dev/badge/skills/hamzabellouch/agent-skills/agent-platform-eval-flywheel.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00108 | $0.03906 |
| Opus 5 | $0.00054 | $0.01953 |
| Sonnet 5 | $0.00022 | $0.00781 |
| Haiku 4.5 | $0.00011 | $0.00391 |
Grade A, and why
agent-platform-eval-flywheel scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 380 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Platform Eval Flywheel Skill
Help users evaluate and iteratively improve GenAI models and agents using
the Agent Platform GenAI Evaluation SDK (google.genai / agentplatform).
When to use this skill
- Evaluating GenAI agents or models with the Agent Platform GenAI
Evaluation SDK (
client.evals.evaluate()). - Creating evaluation datasets from session traces, pandas DataFrames, or synthetic generation.
- Selecting, configuring, or writing custom evaluation metrics.
- Analyzing rubric verdicts, loss patterns, and clustering failures.
- Suggesting concrete code/prompt improvements based on eval results.
- Evaluating a model served on an Agent Platform endpoint (BYOM) or a
Model-as-a-Service (MaaS) model by ID — including deploying the model
first if needed. For this case, follow
references/deployment.md and use the
endpoint_evaluation.py/maas_evaluation.pyscripts.
Safety & Confirmation Tiers (CRITICAL)
Before executing any commands or scripts on behalf of the user, you MUST adhere to the following safety tiers based on the action requested:
- Tier R: Read-only (
inspect_results.py,compare_results.py,validate_dataset.py,parse_adk_traces.py,render_html_report.py)- Rule: No confirmation needed. You may execute these helper scripts immediately to inspect data, validate schemas, parse traces, or compare evaluation results.
- Tier M: Read-only with Compute Costs (
client.evals.run_inference,client.evals.evaluate,client.evals.generate_user_scenarios,client.evals.generate_loss_clusters)- Rule: These operations invoke LLMs or remote evaluation services that consume compute resources and incur costs. This requires interactive confirmation with 'Yes'/'No' options. Once granted once, you do not have to prompt for future evaluation.
Setup
Install the SDK:
pip install google-cloud-aiplatform[evaluation]>=1.154.0 google-genai>=1.0.0
What ships with it
12 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/dataset_schema.md 9.1 KB
- references/deployment.md 6.8 KB
- references/failure_patterns.md 6.2 KB
- references/metric_registry.md 12 KB
- references/sdk_patterns.md 8.0 KB
- scripts/compare_results.py 6.2 KB runs code
- scripts/endpoint_evaluation.py 5.3 KB runs code
- scripts/inspect_results.py 6.9 KB runs code
- scripts/maas_evaluation.py 2.3 KB runs code
- scripts/parse_adk_traces.py 8.4 KB runs code
- scripts/render_html_report.py 3.5 KB runs code
- scripts/validate_dataset.py 11 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 380 lines · 108 tokens per session scan A 265e3208740b
agent-platform-eval-flywheel is a skill published in the GitHub repository hamzabellouch/agent-skills (4 stars, last pushed 1mo ago), licensed MIT. It adds 108 tokens to every session and 3,906 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
spec-implement
Orchestrate parallel implementation of a feature specification by dispatching coder agents batch-by-batch with code review gates between batches. Use this skill when the user says "implement this feature", "start implementing", "run the spec", "execute the plan", "continue implementing", or wants to begin coding a…
spec-create
Create a structured feature specification with self-contained task files organized into parallel execution batches. Use this skill when the user says "create a spec", "plan this feature", "write up an implementation plan", "break this into tasks", or after any planning conversation where the user wants to capture…
spec-ship
Push the current branch to GitHub and create a pull request. Use this skill when the user says "ship it", "ship this", "push to github", "create a pr", "open a pull request", "send for review", "get this reviewed", or wants to push their work and open a PR. Also use when the user says "/spec-ship" (or their host's…
spec-checkpoint
Create a comprehensive checkpoint commit with detailed analysis of all changes. Use this skill when the user says "checkpoint", "commit everything", "save my progress", "create a commit", or wants to stage and commit all current changes with a well-crafted message. Also use when the user says "/spec-checkpoint" (or…
generative-seo
Run an evidence-based SEO + GEO (Generative Engine Optimization) program for ANY website or product codebase — auditing technical SEO, writing or retrofitting content so it gets cited by ChatGPT/Perplexity/AI Overviews, tracking AI-citation visibility, refreshing competitor research, and drafting distribution posts.…
spec-verify
Verify a completed feature spec by driving the real running app — not just lint/typecheck/build — and checking off every acceptance criterion from its specs/{feature}/ folder one by one. Use this after /spec-implement finishes a spec, when the user says "test this end to end", "verify this feature actually works"…