agentleFS
ashrafiucseashrafiucse/security-audit-skills/skills/dotnet-security/SKILL.md

dotnet-security

Skill2 starsChanged 49 days agoMITView on GitHub

Audits ASP.NET Core / .NET Framework applications — Razor Html.Raw and Blazor MarkupString XSS, EF Core FromSqlRaw/SqlQueryRaw and Dapper/ADO string-concat SQL injection, BinaryFormatter/ObjectStateFormatter deserialization RCE, hardcoded machineKey → ViewState RCE, XXE via DtdProcessing.Parse, Newtonsoft TypeNameHandling polymorphic gadgets, AllowAnonymous on sensitive endpoints, antiforgery gaps, CORS AllowAnyOrigin+AllowCredentials, open redirect, Process.Start command injection, hardcoded connection strings, developer exception page in prod. Use when the project has .csproj, .sln, packages.lock.json, packages.config, or web.config.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

.NET Security (ASP.NET Core / .NET Framework)

Step 0 — Detect

rg --files -g '*.csproj' -g '*.sln' -g 'packages.lock.json' -g 'packages.config' -g 'web.config' -g 'appsettings*.json' | head -10
rg -n "Microsoft.AspNetCore|System.Web|TargetFramework" -g '*.csproj'

Record: TargetFramework (netcoreapp/net5+ = ASP.NET Core; net4x = Web Forms/MVC5), template engine (Razor .cshtml, Blazor .razor, Web Forms .aspx), data access (EF Core, Dapper, ADO.NET). Check NuGet versions against ../cve-research/vuln-db/entries/.

Step 1 — Config & secrets

rg -n "Password=|password\"|machineKey|validationKey|UseDeveloperExceptionPage" -g 'appsettings*.json' -g '*.config' -g 'launchSettings.json' -g 'Program.cs' -g 'Startup.cs'
  • Connection strings with embedded passwords ("Password=...") → Critical in public repos, High otherwise
  • <machineKey validationKey="<literal>" decryptionKey="<literal>"/> in web.config → Critical: leaked/hardcoded machine key → forged __VIEWSTATE → pre-auth RCE (KEV-class exploit chains). AutoGenerate is the safe form
  • UseDeveloperExceptionPage() without environment guard, or ASPNETCORE_ENVIRONMENT: Development committed → High (stacktrace + path leak)

Step 2 — Authorization & antiforgery

rg -n "AllowAnonymous|Authorize|ValidateAntiForgeryToken|IgnoreAntiforgery" -g '*.cs'
rg -n "AllowAnyOrigin|AllowCredentials|AddCors" -g '*.cs'
  • [AllowAnonymous] on any action touching data, users, or admin ops → Critical (route-census discipline: disposition EVERY controller/action, ../auth-review/SKILL.md §1)
  • Controller classes with role-sensitive ops and no [Authorize] → High (verify per-handler)
  • Cookie-auth POST actions without [ValidateAntiForgeryToken] → High (state change + CSRF)
  • .AllowAnyOrigin() combined with .AllowCredentials() → High (reflected-origin patterns)
  • Mass assignment: request-bound model assigning privilege fields (user.Role = model.Role) → Critical (CWE-915; the Umbraco GHSL class)

Step 3 — SQL injection

rg -n "FromSqlRaw|SqlQueryRaw|ExecuteSqlRaw|\.Query\(|SqlCommand\(" -g '*.cs'
  • EF Core FromSqlRaw($"... {user}") / ExecuteSqlRaw($"...") with interpolated strings → Critical. Near-miss: FromSqlInterpolated($"... {user}") is SAFE — the API parameterizes C# interpolation by design
  • Dapper conn.Query("... " + user) or $"... {user}" → Critical; conn.Query(sql, new { id }) → safe
  • ADO.NET new SqlCommand("... " + user) → Critical; Parameters.Add/SqlParameter → safe

Step 4 — XSS / templates

rg -n "Html\.Raw|MarkupString|<%=" -g '*.cshtml' -g '*.razor' -g '*.aspx' -g '*.ascx'
rg -n "Mode=\"PassThrough\"|Literal" -g '*.aspx' -g '*.ascx' -g '*.cs'

Census, don't sample: disposition every raw sink with privilege direction (../injection-flaws/SKILL.md XSS table).

  • Razor @Html.Raw(userContent) → direction-triaged (plain @user auto-encodes ✓)
  • Blazor @((MarkupString)userContent) → direction-triaged
  • Web Forms <%= user %> (raw) vs <%: user %> (encoded ✓); <asp:Literal Mode="PassThrough"> → unencoded

Step 5 — Deserialization & XML

rg -n "BinaryFormatter|ObjectStateFormatter|LosFormatter|NetDataContractSerializer|TypeNameHandling" -g '*.cs'
rg -n "DtdProcessing\.Parse|ProhibitDtd\s*=\s*false|XmlResolver" -g '*.cs'
  • BinaryFormatter.Deserialize / ObjectStateFormatter / LosFormatter / NetDataContractSerializer on request/external data → Critical (gadget chains)
  • Newtonsoft TypeNameHandling.All/Auto/Objects enabled on request bodies → Critical (polymorphic gadgets; same class as Jackson defaultTyping). TypeNameHandling.None safe
  • XmlSerializer over type-controlled input → High; XXE: DtdProcessing.Parse or XmlResolver non-null on untrusted XML → High

Step 6 — Files, commands, redirects

rg -n "Path\.Combine|PhysicalFile\(" -g '*.cs' | rg -i "request|query|param|route|input"
rg -n "Process\.Start|ProcessStartInfo" -g '*.cs'
rg -n "Redirect\(" -g '*.cs' | rg -i "request|query|param|url"
  • Path.Combine(root, userInput) serving/reading files → traversal, High
  • Process.Start/ProcessStartInfo with concatenated single-string args → High (argument injection; ArgumentList additive form → safe)
  • return Redirect(Request.Query["url"]) unvalidated → open redirect, High; Url.IsLocalUrl allowlist check → safe

Reporting

Severity table above; fixes localized (bound params, @ encoding, [Authorize] per-route, AutoGenerate machineKey, JSON over BinaryFormatter). Cross-reference NuGet CVEs via ../cve-research/SKILL.md and dependency reachability via ../dependency-vulns/SKILL.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.