fnl-setup-instructions

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's primary function is to manage configuration files for AI agents. While it involves reading and writing to sensitive locations (e.g., ~/.claude/CLAUDE.md, ~/.ssh or .env are mentioned as files to avoid reading), it follows best practices for security and user consent.
  • [COMMAND_EXECUTION]: The skill checks for the presence of tools like claude, codex, and agy on the system PATH to determine which hosts to configure. This is a standard discovery mechanism.
  • [CREDENTIALS_UNSAFE]: The skill explicitly includes a section in its contract template (references/contract-template.md) instructing agents to never read, print, or embed credentials and to avoid files like .env. This promotes secure secret management.
  • [DATA_EXFILTRATION]: No network operations or data exfiltration patterns were detected. The skill operates locally on the user's filesystem and focuses on configuration management.
  • [PROMPT_INJECTION]: The skill does not contain instructions to bypass safety filters or override system prompts. It uses a structured interview process to collect user preferences for "working contracts."
  • [INDIRECT_PROMPT_INJECTION]: While the skill reads existing instruction files (untrusted data), it treats them as drafts for a new configuration and requires explicit user review and confirmation before writing any changes. It defines clear boundaries between user-scoped and project-scoped data.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 03:19 PM