pyright-lsp — agentic threat model
The pyright-lsp plugin presents a low-to-moderate agentic risk, primarily acting as a read-only static analysis tool for Python codebases. The main security vectors involve potential exposure of sensitive source code to the model and vulnerabilities within the local LSP execution environment.
OWASP AIVSS score rationale
| Autonomy of Action | 0.20 | |
| Goal-Driven Planning | 0.10 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.30 | |
| Persistent Memory | 0.10 | |
| Contextual Awareness | 0.50 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| Non-Determinism | 0.20 | |
| Opacity & Reflexivity | 0.20 |
Scored with the canonical OWASP AIVSS formula (AIVSS calculator reference); agentic risk factors estimated from the agent’s described capabilities.
MAESTRO 7-layer threat model
Per-layer threats for this agent. Layers tagged “not certain from listing” are general, caveated commentary where the public description didn’t pin that layer.
Not certain from the listing — relies on Claude's underlying foundation model which is subject to prompt injection, but the plugin itself just provides LSP capabilities to the model.
The plugin directly accesses and processes local Python (.py) files to provide type checking and code intelligence, introducing risks of source code exposure or data exfiltration if the model is compromised.
Integrates via the lspServers config. Risks include tool misuse where the model is tricked into analyzing malicious files, or exploiting vulnerabilities within the Pyright LSP server implementation itself.
Not certain from the listing — the hosting and execution environment of the LSP server (e.g., whether it runs in a sandboxed container or has direct local filesystem access) is not specified.
Not certain from the listing — there is no mention of logging, telemetry, or guardrails to monitor the LSP queries and responses for anomalous behavior.
Not certain from the listing — compliance controls, access policies, or authentication mechanisms for the LSP server are not detailed.
Operates as a single-purpose utility plugin within the Claude ecosystem; there is no indication of multi-agent coordination or delegation capabilities.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).