What it solves
A parser or other risky block can fail without preserving useful diagnostic context.
How it works
1Risky code→2Guard Block→3evidence→4policy action
Use it when
Wrap a checked parser operation
Capture failure evidence before policy handling
Exercise timeout behavior in the host demo path
Expose diagnostic state through the shell helper
Quick Start
#include "loxguard.h"
/* Initialize a guard with caller policy and evidence storage. */
/* Execute the checked block and inspect its explicit result. */
/* Expected: success, or evidence plus a configured policy action. */Engineering evidence
- Repository includes tests, examples and host-only checked-parser demos
- Public and experimental surfaces are separated in docs/API.md
Known limits
- Not a process sandbox or memory-safety proof
- Host-only demo APIs are explicitly demo/test-only
- Platform action hooks and policy are application responsibilities