Praesidia
PassAudited by ClawScan on May 10, 2026.
Overview
Praesidia is a coherent instruction-only integration for agent verification and guardrail management, but it uses an API key, can change guardrail settings, and may send user-provided content to Praesidia.
Install only if you intend OpenClaw to use your Praesidia account for agent verification, discovery, and guardrail management. Use a scoped API key, confirm any guardrail changes before applying them, and avoid sending sensitive content for validation unless you are comfortable sharing it with Praesidia.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Anyone using the skill should assume the assistant can act within whatever Praesidia permissions the API key grants.
The skill requires a Praesidia API key and uses it as the primary credential for authenticated requests.
metadata: {"openclaw":{"requires":{"env":["PRAESIDIA_API_KEY"]},"primaryEnv":"PRAESIDIA_API_KEY"Use a least-privilege Praesidia API key, rotate it if exposed, and avoid installing the skill where untrusted users can trigger actions with that key.
A user-directed request can result in persistent security or compliance policy changes for an agent.
The skill documents an authenticated POST request that creates guardrail configurations for an organization or agent.
POST ${PRAESIDIA_API_URL}/organizations/{orgId}/guardrails ... "Use when:" User asks "Add PII detection" or "Apply toxic language filter"Before applying a guardrail, confirm the organization, agent ID, template, action, scope, and whether the change should affect production.
Messages or documents submitted for validation may be transmitted to Praesidia and processed under that service's data handling policies.
The validation workflow sends user-provided content to Praesidia's external API for guardrail checking.
POST ${PRAESIDIA_API_URL}/organizations/{orgId}/guardrails/validate ... "content": "Text to validate"Do not submit confidential, regulated, or third-party content for validation unless that sharing is intended and covered by the user's Praesidia agreement.
Users may over-rely on a trust score as a guarantee of safety, even though verification is only one input to a security decision.
The response guidance encourages strong safety claims based on Praesidia trust scores.
Recommendation: This agent is fully verified and safe for production use.
Treat trust scores as advisory and combine them with the user's own risk review, permissions checks, and monitoring.