Praesidia

v0.1.0

Verify AI agents, check trust scores (0-100), fetch A2A agent cards, discover marketplace agents, apply guardrails for security and compliance. Use when user mentions agent verification, trust scores, agent discovery, A2A protocol, agent identity, agent marketplace, guardrails, security policies, content moderation, or asks "is this agent safe?" or "find agents that can [task]" or "apply guardrails to protect my agent".

1· 1.3k·0 current·0 all-time
by@msoica·duplicate of @msoica/praesidia
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
Name/description (agent verification, trust scores, guardrails) match the runtime actions: all instructions call Praesidia REST endpoints using a Praesidia API key. Requesting a PRAESIDIA_API_KEY is proportionate for these capabilities.
!
Instruction Scope
The SKILL.md tells the agent to call Praesidia endpoints and to 'get the user's organization ID from their profile or context' but does not specify how to obtain orgId or what contexts are acceptable. This grants broad discretion to the agent and is vague. The instructions also reference PRAESIDIA_API_URL and show exact URL usage; they will transmit user/agent identifiers and content to an external service (Praesidia) when invoked. The combination of vague orgId access and external calls is a scope concern.
Install Mechanism
Instruction-only skill with no install spec, no code files, and no downloads — lowest-risk installation footprint.
!
Credentials
Declared required env var is only PRAESIDIA_API_KEY (primary credential), which is appropriate. However, the SKILL.md and README repeatedly reference PRAESIDIA_API_URL (and local dev URLs) while requires.env does not declare PRAESIDIA_API_URL. The instructions therefore access an environment/config value that isn't listed in the skill metadata — an incoherence that can lead to misconfiguration or unexpected behavior.
Persistence & Privilege
always is false and there's no install-time modification of other skills or system-wide settings. The skill does instruct storing the API key in ~/.openclaw/openclaw.json (typical for OpenClaw skills) which is expected and scoped to the agent.
What to consider before installing
This skill appears to implement a Praesidia integration (it will call praesidia.ai endpoints) and only asks for a PRAESIDIA_API_KEY — that part is reasonable. Before installing, double-check the following: - PRAESIDIA_API_URL mismatch: The documentation uses PRAESIDIA_API_URL (and local dev URLs) but the skill metadata only declares PRAESIDIA_API_KEY. Ensure you set PRAESIDIA_API_URL where required (or confirm the platform supplies a default) to avoid accidental calls to an unexpected endpoint. - API key handling: The skill expects you to store an API key in ~/.openclaw/openclaw.json. Use a least-privilege key if Praesidia supports scoped keys, and avoid exposing long-lived production keys in shared environments. - Data exposure: The skill sends agent IDs, guardrail content, and arbitrary text to Praesidia's API. Review Praesidia's privacy/security policies to ensure sending that data is acceptable for your use case. - orgId ambiguity: The runtime instructions say to obtain orgId from the 'user's profile or context' but do not constrain where to look. Decide and document where orgId should come from (explicit user input, authenticated account profile, or config) to avoid the agent searching unexpected files or contexts for it. - Test in dev first: Use a restricted test API key and the local dev URL option to verify behavior before enabling in production. Monitor the first few requests to confirm that only expected data is transmitted. If you cannot confirm PRAESIDIA's vendor identity or the intended config for orgId/PRAESIDIA_API_URL, treat this skill with caution and ask the skill author or vendor for clarifications and a minimal-scope API key before enabling it.

Like a lobster shell, security has layers — review code before you run it.

latestvk97amgcxpqgs9exnskqcrsbchx80mjkn

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Runtime requirements

🛡️ Clawdis
EnvPRAESIDIA_API_KEY
Primary envPRAESIDIA_API_KEY

Comments