Environment Risk Assessment

Environment Risk Assessment

When to use

• Use before enabling autonomous or semi-autonomous execution.
• Use when an agent runs on a developer workstation.
• Use when production credentials or broad mounts may be present.

Inputs required

• runtime description
• mounted paths
• credential sources
• network and approval settings

Workflow

1. Identify whether the runtime is isolated or shared.
2. Review credentials, mounts, and network scope.
3. Flag unsafe mixes such as prod credentials on a laptop or broad writable mounts.
4. Recommend safer boundaries and required approvals.

Output format

• Environment Summary
• High-Risk Conditions
• Recommended Isolation Changes
• Approval Recommendation

Guardrails

• Default to conservative risk ratings when evidence is incomplete.
• Prefer recommend-and-escalate over automated reconfiguration.
• Cite the observed environment signals for each risk call.
• Do not claim an environment is safe without explicit evidence.

Failure modes

• If the runtime description is partial, mark confidence as low.
• If protected paths are writable, require human review.
• If production credentials are detected, escalate immediately.

---
name: environment-risk-assessment
description: Evaluate whether an agent runtime is isolated enough for the
  current task and flag unsafe mixes of credentials, mounts, and privileges.
---