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/VuDZ/RoslynMcpServerWrote 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/vudz/roslynmcpserver/roslyn-mcp-ci)<a href="https://agentmods.dev/rules/vudz/roslynmcpserver/roslyn-mcp-ci"><img src="https://agentmods.dev/badge/rules/vudz/roslynmcpserver/roslyn-mcp-ci/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/vudz/roslynmcpserver/roslyn-mcp-ci"><img src="https://agentmods.dev/badge/rules/vudz/roslynmcpserver/roslyn-mcp-ci.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 | $0.00548 | $0.00548 |
| Opus 5.5 | $0.00219 | $0.00219 |
| Sonnet 5.5 | $0.00110 | $0.00110 |
| Haiku 4.5 | $0.00055 | $0.00055 |
Grade A, and why
roslyn-mcp-ci 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 20d 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 — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CI test jobs (this repo)
Canon: .github/workflows/test-suite.yml.
ci.yml and release.yml call that workflow. Do not copy a second dotnet test matrix.
The repo is public — GitHub-hosted minutes are not the limit (~20 concurrent jobs are).
Do not fold CI back into one sequential dotnet test on the whole .sln to “save runners”.
What may not run in one testhost
[CollectionDefinition("AnalyzerLifecycle", DisableParallelization = true)] plus
AnalyzerShadowPublisher: one MSBuild lifecycle host / shadow dirs. In-process
xUnit threads will race. Unit tests (Category!=AnalyzerLifecycle) already parallelize
inside the testhost.
Partition contract
Union of filters must be every discovered test; the two jobs must not overlap.
| Job | Filter | ROSLYN_MCP_ANALYZER_LIFECYCLE |
|---|---|---|
unit |
Category!=AnalyzerLifecycle |
unset |
lifecycle |
Category=AnalyzerLifecycle |
1 |
AnalyzerLifecycleHostTests is the only class carrying Category=AnalyzerLifecycle
(thirteen host methods). Epoch1SemanticInventoryTests is source-only, carries no trait, and
runs in unit. Do not add named per-epoch shards back, and do not leave an empty named shard.
AnalyzerLifecycleFactAttribute / AnalyzerLifecycleTheoryAttribute read the variable before
the host probe: anything other than 1 skips without starting an MSBuild host. MCP tools never set
it, so an agent must not pick these methods through run_specific_test — the run would skip.
fail-fast: false. Skip is a job failure (.github/scripts/Assert-NoSkippedTests.ps1), not a
green CI. Never weaken that script.
Pass --filter as a quoted env value ("$env:FILTER"). Bare & in PowerShell is the call operator.
JS actions must be Node 24 majors (checkout@v6, setup-dotnet@v5, upload-artifact@v7,
download-artifact@v8). Do not pin @v4 of those — GitHub removes Node 20. Do not set
ACTIONS_ALLOW_USE_UNSECURE_NODE_VERSION. Hosted windows-latest is new enough.
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.
- 20d ago Changed · -1 lines · -23 tokens per session b689bdfd82d4
- 26d ago First seen · 47 lines · 571 tokens per session scan A 8f726506eaf8
roslyn-mcp-ci is a cursor rule published in the GitHub repository VuDZ/RoslynMcpServer (8 stars, last pushed 11d ago), licensed MIT. It adds 548 tokens to every session, about $0.0022 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 2026-09-22.
Other cursor rules, from other repositories
workflow
BDD-style workflow, UI screenshots in the PR, HTTP OpenAPI, config schema and CLI packaging sync before lint, final checks, the comment on an issue a merge closes and the cleanup after the merge.
build-tags
The shipped build tag set is spelled in several places that move together.
pytest-integration-tests
Below is an example test. Notice the following.
pytest-tests
Below is an example test, you'll notice the following.
ci-cd
CI/CD pipeline patterns and deployment rules.
tooling-for-apps
ArchiPy app tooling — uv dependencies, project checks, Behave, Ruff, ty, and pre-commit.