Reversibility-aware safety signals for LLM tool-use agents | estimating action recoverability before irreversible operations: (delete/send/pay/commit) when rollback is unavailable.

A protocol for recording and preserving decision memory: context, intent, acknowledged risk, reversibility, and outcomes over time.

Blast radius modeling framework for agentic systems — 5 axes (action class, chain depth, external reach, reversibility window, identity scope), OSI cross-walk, and standards alignment.

Autonomous execution governance for coding agents: run a ticket backlog to completion unattended. ASK/PARK/HALT autonomy contract, durable per-ticket state machine, deterministic test gates, git-backed reversibility, heartbeat watchdog, morning report. Python stdlib only.

Essays on agentic AI security, decision-rights, reversibility-graded authority, manifest-declared action class, deterministic gates, and standards contribution method.

This repository offers operational and application design frameworks addressing the specific challenges of public sector using AWS Well-Architected-Tool.

Supplementary materials for Reversibility as a Safety Gate for Agentic Actions

A consent-first framework for human–ASI collaboration: versioned, cryptographically signed, and written from lived experience of identity continuity.

Installable guard skill for agent skills.

SHA256

Archival mirror of the PHRONESIS Corpus Definition (v1.0), a non-binding, non-authoritative philosophical framework on restraint in the age of advanced AI.

Give your AI agent an undo button — it flags which actions are reversible, compensable, or permanent before they run. Try the live demo.

Analysis scripts underlying the assessment of acidification hysteresis in the Arctic Ocean

MCP server for reversibility intelligence — check if actions can be undone

Reference implementation of OWASP AISVS C9.2.6 and C9.2.7: manifest-declared action class, deterministic gate, worst-case chain rule. JSON schema + Python.