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 Rheinmir/setup --skill dym-harness-sdk-effect-testsgit clone --depth 1 https://github.com/Rheinmir/setupWrote 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/rheinmir/setup/dym-harness-sdk-effect-tests)<a href="https://agentmods.dev/skills/rheinmir/setup/dym-harness-sdk-effect-tests"><img src="https://agentmods.dev/badge/skills/rheinmir/setup/dym-harness-sdk-effect-tests/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/rheinmir/setup/dym-harness-sdk-effect-tests"><img src="https://agentmods.dev/badge/skills/rheinmir/setup/dym-harness-sdk-effect-tests.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 | $0.00063 | $0.00758 |
| Opus 5.5 | $0.00025 | $0.00303 |
| Sonnet 5.5 | $0.00013 | $0.00152 |
| Haiku 4.5 | $0.00006 | $0.00076 |
Grade A, and why
dym-harness-sdk-effect-tests 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 yesterday.
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 — 35 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: dym-harness-sdk-effect-tests — test hiệu ứng, không test lời văn
When to use
- Có agent chạy model thật và cần test không vỡ vì model trả lời khác chữ.
- Nguyên tắc gốc (
harness-py/tests_integ/conftest.py:7-8): "Assertions check observable side effects (a file created, a tool blocked), never the model's exact prose — the model is non-deterministic, the effect is not." - Sinh ra: bộ pytest nằm NGOÀI
testpaths(unit run mặc định không gom), chạy tường minh bằngpytest tests_integ.
Steps
- Fixture dựng agent tối giản — model nhỏ, tắt reasoning, tắt plugin, tắt session — để test chỉ đo đúng thứ nó dựng:
INTEG_MODEL = os.environ.get("STRANDS_INTEG_MODEL", "bedrock/global.anthropic.claude-haiku-4-5-20251001-v1:0") def build_agent(**o): return create_harness(**{"model": INTEG_MODEL, "effort": "off", "builtin_plugins": [], "session": False, **o}) - Cwd tạm cho mỗi test (
monkeypatch.chdir(tmp_path)) — tool file/shell của agent phân giải theo cwd. - Ba predicate đọc lịch sử hội thoại (
agent.messages), không đọc câu trả lời cuối:tool_succeeded(agent, name)— cótoolUsetên đó VÀtoolResultcùngtoolUseIdvớistatus == "success".tool_attempted(agent, name)— cótoolUsetên đó bất kể kết quả → chống pass giả khi assert "tool bị chặn" (chứng minh nó đã THỬ rồi bị deny, không phải bị bỏ qua).tool_result_contains(agent, name, needle)— output CỦA TOOL chứaneedle.
- Một lượt model phủ nhiều tool để rẻ (e2e của họ: 7 tool/1 lượt, rồi assert file trên đĩa + kết quả subagent chứa
"333"). - Kiểm cơ chế qua số đo hạ tầng, ví dụ cache:
result.metrics.accumulated_usage["cacheReadInputTokens"]phải > 0 và tăng mọi lượt. - Chạy:
pytest tests_integ --reruns 2(cầnpytest-rerunfailures, credential model).
Rules
- Không assert chuỗi model nói; assert trạng thái (file, tool result, metric).
- Mọi assert "không làm X" phải đi kèm
tool_attempted— nếu không sẽ pass khi model đơn giản là không thử. --reruns 2che độ flaky thật: nếu muốn biết tỉ lệ ổn định thì chạy N lần không rerun và ghi pass@k — repo gốc KHÔNG làm việc này.- Chưa kiểm chứng ở đây: 11 test integ chỉ được
--collect-only(rc 0), chưa chạy vì cần AWS Bedrock.
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.
- yesterday First seen · 35 lines · 63 tokens per session scan A 49dc85d8d3f2
dym-harness-sdk-effect-tests is a skill published in the GitHub repository Rheinmir/setup (11 stars, last pushed 20d ago), licensed MIT. It adds 63 tokens to every session and 758 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-11-07.
dym-harness-sdk-effect-tests compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| dym-harness-sdk-effect-testsRheinmir/setup | Skill | 11 | 20d ago | MIT | A | 63 | Claude Code |
| verify-omnigentomnigent-ai/omnigent | Skill | 11k | today | Apache-2.0 | A | 100 | Claude Code, Codex |
| test-masterJeffallan/claude-skills | Skill | 12k | 8d ago | MIT | A | 104 | Claude Code |
| eval-dataset-designbojieli/ai-agent-book | Skill | 57k | 1mo ago | Apache-2.0 | A | 138 | Claude Code, Codex |
| dogfoodcallstack/agent-device | Skill | 5.3k | today | MIT | A | 55 | Claude Code |
| cua-driverdavidondrej/skills | Skill | 4.3k | 2d ago | MIT | A | 68 | Claude Code, Codex |
| test-warp-uiwarpdotdev/warp | Skill | 66k | today | AGPL-3.0 | A | 55 | Claude Code |