cli-build

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends using cligentic from a GitHub repository (Railly/cligentic) to provide standardized CLI infrastructure blocks.
  • [REMOTE_CODE_EXECUTION]: Suggests installing the cligentic skill via the npx skills add command. The skill incorporates a safety protocol by instructing the agent to wait for user consent before installation and provides a manual alternative if consent is withheld.
  • [INDIRECT_PROMPT_INJECTION]: Proactively addresses the risk of indirect prompt injection from third-party APIs. Ingestion points: Third-party API response text as described in SKILL.md and Phase 4. Boundary markers: Recommends using versioned JSON envelopes (json-contract.md). Capability inventory: Mutating commands, network calls, and file writes. Sanitization: Explicitly instructs developers to treat external content as untrusted and escape it (Phase 4).
  • [COMMAND_EXECUTION]: Instructs the use of standard development environment tools such as bun link and npm link to verify the installation and behavior of built CLIs in a production-like setting.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 03:45 AM
Security Audit — agent-trust-hub — cli-build