exa-policy-guardrails

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is a utility designed to enhance the security posture of projects using Exa. It implements patterns to prevent hardcoded secrets (API keys), enforces TypeScript type safety, and provides templates for Policy-as-Code (OPA) and CI/CD checks.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface as it is designed to read and analyze external project files.
  • Ingestion points: Reads source code files (TS/JS) and JSON configuration files (SKILL.md).
  • Boundary markers: None; the skill does not define specific delimiters to separate code-under-analysis from agent instructions.
  • Capability inventory: The skill has permissions for Read, Write, Edit, and Bash(npx:*), which could be exploited if an attacker-controlled file contains instructions the agent obeys.
  • Sanitization: There is no explicit sanitization or filtering mechanism for the content of files ingested during linting or policy checks.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 01:02 AM