Tool-call guardrails for Claude Code derived from cc-logger telemetry: nudge recoverable mistakes, block irreversible ones. The guard stage of callusguard; still works standalone.
The PSA Control Standard defines governance controls for deployed agentic systems.
Tool-call guardrails for the OpenAI Codex CLI: nudge recoverable mistakes, block irreversible ones, keep an audit trail. The Codex guard stage of callusguard; still works standalone.
A small reference implementation for controlled execution of AI agent actions.
Operational AI-human interaction controls focused on escalation visibility, intervention integrity, and behavioral oversight.
Plan a governed maker action across the Loomground family, then control the maker over an A2A protocol. Two profiles: a protocol floor and a fail-closed Loomground compliance team.
Egress and ingress locks with capability tokens, and at-rest sealing of a folder's memory.
Deterministic evaluation harness for AI agent runtime controls, mutation testing, and decision/effect integrity against a pinned SUT.
Leased autonomy that decays unless renewed, tripwires that quarantine, drift measured against a baseline.
Evidence-backed agent runtime controls for bounded tool execution: scoped capabilities, deterministic ALLOW/DENY/ASK decisions, tool-bound human approval, result gating, adversarial evaluation and authority incident evidence. Scoped ACS v0.1.0 reference implementation.
Span-by-span PII-cleaned overlay and egress guard; only the overlay leaves the machine
Controller-signed erasure sweeps, pending-erase markers, forgotten-subjects ledger
100% Generic Multi-Environment Logger with Advanced Configuration - Smart detection, file-level overrides, and beautiful console formatting
Durable approval envelopes for governed agents: what a grade covers, checked each iteration, appended to the signed audit chain.
To associate your repository with the runtime-controls topic, visit your repo's landing page and select "manage topics."