القائمة
Run and interpret opt-in expensive validation lanes: full slow tests, IL round-trip sweep, corpus sensors, validity scans, and analysis census.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
Find evidence instead of guessing for .NET packages, platform libraries, local assemblies, APIs, dependencies, and version-to-version API changes.
Whole-assembly call-graph leverage ranking and performance triage for libraries (experimental).
Decide whether a change or upgrade is safe — API-surface diffs, behavioral diffs (allocations, exceptions), feature switches, and version resolution.
Judge whether code is sound and safe to call — its exception surface (throws, catches, exception types) and the unsafe operations in a method body.
Reconstruct a method or type locally as C# and IL — decompiled source, annotated source with hidden facts, raw IL, fidelity levels, and IL-offset lookup. Always local, no network.
Use for opt-in Deep Inspect run triage: test lane failures, census artifacts, pinned-subset drift, and routing findings into issues or tracker updates.
استنادا إلى تصنيف SOC المهني
| name | dotnet-inspect-deep-inspect |
| description | Run and interpret opt-in expensive validation lanes: full slow tests, IL round-trip sweep, corpus sensors, validity scans, and analysis census. |
Use this skill when a change needs expensive evidence outside normal PR CI. Deep Inspect is opt-in: run it manually when preparing risky PRs, and the test lane also runs during publish before packages are built.
| Lane | Use for | Runs |
|---|---|---|
test | Blocking proof before publish or risky merges | Full decompiler tests, full analysis tests, vendored ILAssembler restore, full IL round-trip sweep. |
census | Observational broad signal and triage | Real-world corpus sensor, validity predicate scan, uncapped validity sweep, assertion scan, analysis corpus sensor, paydirt recall. |
all | Release-candidate deep read | Both lanes. |
Run manually:
gh workflow run deep-inspect.yml -f lane=test
gh workflow run deep-inspect.yml -f lane=census
gh workflow run deep-inspect.yml -f lane=all
Inspect recent runs and artifacts:
gh run list --workflow deep-inspect.yml --limit 10
gh run view <run-id> --log-failed
gh run download <run-id> -D /tmp/deep-inspect-<run-id>
For the test lane:
dotnet run --project src/ILInspector.Decompiler.Tests -c Release
dotnet run --project src/ILInspector.Analysis.Tests -c Release
bash eng/restore-ilassembler.sh
dotnet run --project tests/DotnetInspector.ILRoundtrip.Tests -c Release
For the fast PR subsets:
dotnet run --project src/ILInspector.Decompiler.Tests -c Release -- -trait- "Speed=Slow"
dotnet run --project src/ILInspector.Analysis.Tests -c Release -- -trait- "Speed=Slow"
dotnet run --project tests/DotnetInspector.ILRoundtrip.Tests -c Release -- -trait- "Speed=Slow"
For the census lane, prefer the workflow so artifacts are retained. If running
locally, use the same scripts/baselines as deep-inspect.yml and preserve the
generated snapshots/cards under /tmp or artifacts/ for review.
test failures as blockers: reproduce locally, identify the first
failing proof, and fix or explicitly defer before publish.census output as triage signal unless a command exits nonzero by
design. Compare snapshots against committed baselines and route meaningful
drift to issues or follow-up PRs.[Trait("Speed", "Slow")] and keep them in Deep Inspect / publish / full
local runs.