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.
git clone --depth 1 https://github.com/sweeden-ttu/canvas-lms-mcpWrote 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/rules/sweeden-ttu/canvas-lms-mcp/no-synthetic-data)<a href="https://agentmods.dev/rules/sweeden-ttu/canvas-lms-mcp/no-synthetic-data"><img src="https://agentmods.dev/badge/rules/sweeden-ttu/canvas-lms-mcp/no-synthetic-data/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/rules/sweeden-ttu/canvas-lms-mcp/no-synthetic-data"><img src="https://agentmods.dev/badge/rules/sweeden-ttu/canvas-lms-mcp/no-synthetic-data.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 | $0.00824 | $0.00824 |
| Opus 5 | $0.00412 | $0.00412 |
| Sonnet 5 | $0.00165 | $0.00165 |
| Haiku 4.5 | $0.00082 | $0.00082 |
Grade A, and why
no-synthetic-data 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 11d 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 — 54 lines — stays where its author put it; the contents beside it link to each section on GitHub.
No Synthetic, Mock, or Dummy Data (Global Rule)
Scope
This rule applies to all work in this project: evidence evaluators, examples, templates, presentations, tests, documentation, and any code or content that could introduce fake data.
Prohibited
- Synthetic data: Artificially generated data that is not from a real source (e.g. fabricated API responses, generated user records).
- Mock data: Placeholder or fake data used to simulate behavior (e.g. mock objects returning hardcoded values, mock APIs with fake JSON).
- Dummy data: Filler data (e.g. "foo", "[email protected]", "Sample User") used in examples, templates, or presentations instead of real or experiment-derived data.
Do not use synthetic, mock, or dummy data in evidence evaluators, examples, templates, presentations, or tests. If you find such data, do not leave it in place; replace it using the process below.
Required alternative: Hypothesis, experiment, evidence
Whenever fake data would be introduced:
- Hypothesis: State a testable hypothesis for what real data or behavior is needed (e.g. "Canvas API returns assignments in this shape").
- Planned experiment: Define a concrete experiment to obtain or validate real data (e.g. call the live Canvas API, fetch real course content, crawl real documentation).
- Evaluate evidence and results: Run the experiment, collect evidence, and update beliefs (e.g. via BayesianOrchestrator or EvidenceEvaluatorAgent). Use the resulting real data in examples, templates, and presentations.
Prefer live/verified endpoints, real course IDs from test_hints.json, and real documentation URLs. If real data cannot be obtained (e.g. no API key), document the gap and the minimal experiment that would be run; do not substitute mock/synthetic/dummy data.
Required: review-changes step
Every change set (PR, commit batch, or delivered task) must include a review-changes step that:
- Evaluates step-by-step instructions: Review the instructions that were followed and confirm they do not rely on synthetic/mock/dummy data.
- Performs peer review: Apply a peer review (e.g. via the cs-peer-reviewer-trustworthy-ai skill or equivalent) to the changes.
- Attempts reproduction: Run the same steps in a clean environment and verify outcomes (e.g. re-run tests, re-fetch data, re-build presentation).
- Accept or reject premise:
- Accept: If reproduction succeeds and no synthetic/mock/dummy data was introduced, accept the premise and keep the change.
- Reject: If mock/synthetic/dummy data would have been used or results cannot be reproduced, reject the premise and revert or rewrite the change so that real data and reproducible evidence are used instead.
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.
- 11d ago First seen · 54 lines · 824 tokens per session scan A d730728daeb9
no-synthetic-data is a cursor rule published in the GitHub repository sweeden-ttu/canvas-lms-mcp (0 stars, last pushed 6mo ago), licensed MIT. It adds 824 tokens to every session, about $0.0041 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 cursor rules, from other repositories
20-testing-generated-and-structure
Testing expectations, generated-file boundaries, and CLI structure contract.
fuzzing
Fuzzing setup and guidelines for skia-rs.
testing-and-quality
The testcomprehensive.py provides systematic testing of all 31 MCP tools.
smoke-monkey-tester
Smoke tests for core functionality and monkey/chaos tests aiming to break the system with randomized inputs. / TR: Sistemin temel fonksiyonlarını kontrol eden smoke testler ve rastgele girdilerle sistemi çökertmeyi hedefleyen monkey/chaos testleri.
STAGE_S5_TEST_POINTS
A staged workflow for turning approved requirements into a complete list of test points. It checks required earlier documents and records any requirement that has no matching test point.
testing
Testing conventions for Ion Design System using Jest and Angular Testing Library.