The runtime layer that watches a live agent and can interrupt it: deciding which actions need a gate before they run, the interception points where a check can actually stop something, output screening before a user or system receives it, the interrupt harness and how interruption itself fails, blast-radius bounds, incident response, and the written record. It explains how to build the enforcement layer around a running agent; it never declares an agent safe and never recommends a moderation vendor. Use when asking "agent guardrails", "agent safety layer", "action interception", "human in the loop gate", "output filtering for agents", "kill switch for agents", "agent incident", "irreversible action gate", "prompt firewall", "agent guardrail design". Not for tool permissioning at build time (ac-security-engineer), not for analyzing injection attacks as a vulnerability class (ac-prompt-engineer), not for content adjudication policy (ac-content-moderation), not for service-level error budgets (ac-reliability).
cp -r cryptoskill/skills/exchanges/anoscranel-ac-agent-guardrails .claude/skills/clawhub install anoscranel-ac-agent-guardrailsCapabilities below are detected automatically by an open-source scanner that reads the skill's text and scripts (see how this is computed). Not measured means the scanner couldn't make a confident call — it is NOT a green check, and you should treat it as a possible red flag until a human or a stronger scanner has measured it.
unknown
We did not find any well-known hosted services in this skill's text or scripts. This does NOT mean the skill is local-only — it might use services we don't yet recognize, or talk to them through code paths our scanner can't reach. A complete dependency list (every package, library, and binary, with integrity hashes) is on the roadmap; today we only show recognized hosts.
No one has audited this skill yet. That is different from “audited and clean” — it just means no professional reviewer (a security firm, the CryptoSkill team, or a verified independent researcher) has signed off on it. There are no audit reports to read. How reviewer levels work →
SKILL.md → SOURCE.md → TRUST.auto.yaml → Browse directory →
Auto-generated by cryptoskill/extract-capabilities/0.3.1 · hosted-service list version 2026-09-06 · how this is computed