feat: add script-based rule evaluation

.claude/commands/ci.md

@@ -0,0 +1,7 @@
+Commit working changes and push them to CI
+
+Run `cargo clippy -- -D warning` before pushing changes.
+
+Enter loop where you wait for CI to complete, resolve issues,
+and return to user once CI is green or a major decision is needed
+to resolve it.

.claude/settings.json

@@ -7,10 +7,6 @@
           {
             "type": "command",
             "command": "cargo fmt"
-          },
-          {
-            "type": "command",
-            "command": "cargo clippy --all-targets -- -D warnings"
           }
         ]
       }

.github/workflows/tests.yml

@@ -41,6 +41,9 @@ jobs:
       - name: Run smoke tests
         run: cargo nextest run --profile ci --test smoke_test --verbose
 
+      - name: Run script integration tests
+        run: cargo nextest run --profile ci --test script_integration --verbose
+
       - name: Run weak mode integration tests
         run: |
           # On macOS, we only support weak mode due to PF limitations
@@ -112,6 +115,11 @@ jobs:
           source ~/.cargo/env
           cargo nextest run --profile ci --test smoke_test --verbose
 
+      - name: Run script integration tests
+        run: |
+          source ~/.cargo/env
+          cargo nextest run --profile ci --test script_integration --verbose
+
       - name: Run Linux jail integration tests
         run: |
           source ~/.cargo/env
@@ -148,6 +156,9 @@ jobs:
       - name: Build
         run: cargo build --verbose
 
+      - name: Run script integration tests
+        run: cargo nextest run --profile ci --test script_integration --verbose
+
       - name: Run weak mode integration tests
         run: cargo nextest run --profile ci --test weak_integration --verbose
 

CLAUDE.md

@@ -51,3 +51,7 @@ on both targets.
 ## Formatting
 
 After modifying code, run `cargo fmt` to ensure consistent formatting before committing changes.
+
+## Logging
+
+In regular operation of the CLI-only jail (non-server mode), info and warn logs are not permitted as they would interfere with the underlying process output. Only use debug level logs for normal operation and error logs for actual errors. The server mode (`--server`) may use info/warn logs as appropriate since it has no underlying process.