Jev 101
Start Guides
Search
中文 EN

use cases tool-call-guardrails

Agents ✦ 94outlineAuthor-reported metrics — re-measure on your hooks

Tool-call / jailbreak guardrails

Who: Agent harness authors. Steps: (1) State = tool name, args summary, user utterance, rules. (2) Prefer one Noul per hazard (destructiveness, exfil, scope creep) — not one vague “is this dangerous?”. (3) Code maps probabilities → allow / ask / block with per-hazard thresholds. Expected effect: fewer false negatives vs a single broad question (firewall/pi-warden authors report this pattern — cite as author-reported).

Source: AnshChoudhary/typesafe-ai-firewall ↗ Docs / source ↗
1

Treat this as an outline — adapt state and questions to your data.

2

Implement in code — call System One / Jev; compose answers yourself.

3

Gate on confidence — act, confirm, or escalate before side effects.

Outline sketch

Prompt 
Audience: agent harness authors.
Steps:
1) Capture proposed tool + args summary + utterance + rules.
2) Parallel Nouls — one per hazard (destructiveness, exfil, scope_creep, injection_in_tool_result).
3) Code: per-hazard thresholds → allow / ask user / block.
Expected effect: stronger screening than one vague danger question (author-reported in community firewall studies).
Do not execute tools from inside the model.

Needs access to: typesafe-sdkagent-runtime

Who it's for

Agent harness authors

Steps / how it's set up

Insert a pre-execution gate. Sources: llm_guardrails cookbook · typesafe-ai-firewall · pi-warden · pi-jev. Measured claims (paired runs / hard-negative rates) are author-reported.

Prefer one Noul per hazard (exfil, irreversible, jailbreak, …) with per-threshold tables in code rather than one vague “is this safe?” question. Multi-hazard firewalls are author-reported improvements over single checks.

Sources (wave-2 deepen)

Expected effect

Guardrails are cheap relative to the tool call they protect; probabilities beat essays.

Unofficial outline for learning. Paraphrased from public docs and tutorials — not a production recipe. Review sources before you automate anything.