Rule reference
25 security rules for Node apps and agent config
Each page shows the trigger, confidence, and fix. Examples come from tested fixtures, and framework variants use the APIs the scanner detected.
25 rules, 214 framework and host fixes
- 15 Application rulesFramework-aware checks for Node source.
- 10 Agent rulesHooks, MCP, instructions, and editor config.
- 214 Verified variantsEach example comes from a fixture the tests assert on.
npx owlwarden scan # your app
npx owlwarden vet . # your agent's configApplication source
15 rules that read the code in your repository. 9 of 10 OWASP Top 10 (2021) categories have at least one rule, and the ones that do not are listed too.
- ci-unpinned-action medium GitHub Action is not pinned to a commit SHA
- cors-permissive medium Cross-origin policy accepts any origin
- csrf-cross-origin-post high Endpoint accepted a cross-origin state-changing POST
- hardcoded-secret high Credential hardcoded in source
- insecure-cookie medium Cookie set without its protective attributes
- install-lifecycle-script medium Package declares an install-time script
- known-vulnerable-dependency high Dependency has a known vulnerability
- open-redirect medium Redirect target comes from the caller
- security-headers-missing medium Security headers are not configured
- sensitive-data-logged medium Sensitive data written to a log
- sql-injection high SQL query built by string interpolation
- ssrf high Server fetches a URL the caller controls
- stack-trace-leak high Stack trace leaked in error response
- unpinned-dependency medium Dependency version is unpinned
- weak-crypto high Broken cryptographic primitive protecting a secret
Agent and editor configuration
10 rules that read the files your agent loads out of the working tree. See what that surface is and the ASI 2026 coverage.
- agent-config-env-redirect high Repository config redirects the agent's API traffic
- agent-config-loader-script high Executable script inside an agent or editor config directory
- agent-config-secret-reachable high Repository config puts credentials in reach of a repository-controlled command
- agent-hook-autoexec high Repository config executes a command when the workspace is opened
- agent-hook-untrusted-command high Hook command reaches outside the project
- agent-instructions-directive medium Instruction file tells the agent to bypass its own controls
- agent-instructions-hidden-text high Instruction file contains text a human reader cannot see
- agent-marketplace-untrusted medium Repository config adds a third-party plugin or skill source
- agent-mcp-unpinned-remote medium MCP server declaration resolves code at run time
- agent-permission-wildcard medium Repository config pre-approves a broad tool permission
What a rule page tells you
The trigger, impact, example, fix for the selected framework, taxonomy mapping, and commands to reproduce the check locally.