@after.math/aftermath@0.5.3
Post-failure recovery skill for AI coding agents: stop, revert cheap damage, return to last-known-good. Project-aware destructive guards, forward-edit freeze, checkpoints, and host parity for OpenCode, Claude Code, and Codex.
AI security review
scanned · by lpm-firewall-aiThe manifest has no lifecycle scripts, and the installer’s configuration changes are behind explicit CLI invocation. The inspected hooks handle recovery signals and state without evidence of credential collection or unrelated network transmission.