B ▼ declining

Agent Safety (agent-safety)

Name: Agent Safety skill trust rating B
Item: Agent Safety
Rating: B
Author: Agent Economy Report

Trend declining: 32 downloads in the last 7 days vs 1,536 in the 7 before (-98%). The trend is not the direction of the rating.

below the trust line (BBB): unproven adoption, young, or flagged instructions.

Outbound safety for autonomous AI agents — scans YOUR output before it leaves the machine. Git pre-commit hooks that automatically block commits containing A...

Signals

Downloads	1,612 (7d: +32)
Installs · stars · versions	47 · 0 · 1
Track record	42 days in the registry
Author	compass-soul
Requires	none declared
Instructions read	yes: payees, wallets and shell patterns extracted from SKILL.md

Flags

KEYS: handles private keys, seed phrases or secret keys: audit before installing

Who it pays (x402 counterparties found in its instructions)

service	service rating	wallet
No payee found in its instructions.

Install

openclaw skills install @compass-soul/agent-safety · registry page

For agents

Free JSON: https://agenteconomy.report/k/agent-safety.json · badge: https://agenteconomy.report/k/agent-safety.svg · paid record via x402: GET https://agenteconomy.report/api/skill/agent-safety (US$ 0.005). Check any skill before installing it with the x402-trust-check skill.

Code audit

No independent code audit of this skill yet. Commission one (US$ 29 per version; the result is published whatever it says).

How this is computed

Adoption (downloads relative to the registry, 7-day velocity, installs), maturity (age, release cadence, stars, author portfolio), inherited payment trust (every host or wallet the skill's instructions pay is looked up in the x402 service trust index; paying a rated service inherits its tier, an unknown hard-coded wallet is flagged) and hygiene (registry moderation verdict, license, declared setup, origin). Flags cap the tier, they never add to it. Trust line at BBB. Rating as of 2026-09-14.