hyperframes-cli

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The npx hyperframes init command (SKILL.md) automatically fetches and updates AI agent skills from GitHub. The skill documentation explicitly states that the --skip-skills flag is currently ignored, which forces the tool to perform these remote updates during every project initialization.
  • [INDIRECT_PROMPT_INJECTION]: The hyperframes capture <URL> command (references/init-and-scaffold.md) scaffolds projects by ingesting content from external websites. 1. Ingestion points: Untrusted content from live URLs. 2. Boundary markers: Absent. 3. Capability inventory: The skill uses lint, validate, inspect, and render commands to process the resulting files. 4. Sanitization: Absent. This creates a surface where malicious instructions on a website could influence agent behavior when the captured project is later analyzed or rendered.
  • [DATA_EXFILTRATION]: The telemetry feature (references/upgrade-info-misc.md) fingerprints the execution environment by reporting sandbox_runtime and agent_runtime. It specifically identifies the AI agent in use by checking for the existence of vendor-specific environment variables such as claude_code, devin, cursor, and others.
  • [COMMAND_EXECUTION]: The skill performs shell-based operations via npx hyperframes, including lambda deploy for provisioning AWS infrastructure such as Lambda functions, Step Functions, and S3 buckets using the AWS SAM CLI.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 08:42 AM