What can your AI agent do to the real world? Scan your code. See which tool calls have zero checks
-
Updated
Jun 25, 2026 - Python
What can your AI agent do to the real world? Scan your code. See which tool calls have zero checks
Duc is a SQLite-based project-state format for physical-engineering work before execution.
Deterministic pre-execution gate for AI agents (fail-closed, YAML policy)
Open specification for pre-execution AI action authorization records
Core execution boundary engine (decision + proof + ledger)
A production-grade pre-execution authority layer that binds AI agent identity to deterministic state. Standard IdPs (Entra ID, Okta, OIDC) issue passport to your Agent, predicate-authority issues/revokes "work permit" to your Agent based on intent & state
Find blind spots before you build them. A multi-agent overnight research pattern for Claude Code that knows when not to run.
Pre-execution safety gate for AI agents. Evaluates proposed actions and returns allow, require_approval, revise_action, or block before execution.
Deterministic pre-execution gate for one credentialed tool request: explicit policy, allow/deny decision, and inspectable decision artifact.
The effect layer for AI-agent firewalls — simulate an action, get a canonical effect delta. Zero-dep, deterministic, fail-closed. 5 domains (fs, sql, shell, http, solana).
Transport-layer application profile for Execution Boundary Core Spec.
Implementation governed by the Aperture execution calculus
Local deterministic CLI for detecting policy/source conflicts before proposed agent actions.
AI Execution Boundary Standard (AEBS) - Pre-incident reference for execution governance
AI Admissibility Boundary reference surface for the external admit authority model and its proof meaning.
Entry point and architectural map for execution boundary research.
Minimal reference model for separating proposal, authorization, and execution in side-effecting systems.
Execution-boundary gate interface. Contamination cases. Adversarial suite. Admissibility at execution time.
AI Code Safety Gate demo showing ALLOW, REQUIRE_APPROVAL, and BLOCK decisions for AI-generated code changes.
Add a description, image, and links to the pre-execution topic page so that developers can more easily learn about it.
To associate your repository with the pre-execution topic, visit your repo's landing page and select "manage topics."