Skip to content

docs(ai-devkit-setup): scope unsandboxed execution to listed commands - #380

Closed
codeaholicguy wants to merge 1 commit into
mainfrom
security/ai-devkit-setup-sandbox-scope
Closed

codeaholicguy wants to merge 1 commit into
mainfrom
security/ai-devkit-setup-sandbox-scope

Conversation

@codeaholicguy

@codeaholicguy codeaholicguy commented Oct 14, 2026 •

Copy link
Copy Markdown
Owner

Summary

Hardens the ai-devkit-setup skill per the skill security audit.

  • Finding: skills.sh Gen Agent Trust Hub Warn / MEDIUM — PRIVILEGE_ESCALATION (explicitly directs the agent to run outside filesystem sandboxes to modify host config like ~/.pi/agent), COMMAND_EXECUTION, EXTERNAL_DOWNLOADS (npx @latest fallback).
  • Change: the unsandboxed-execution note now applies only to the ai-devkit status --json / ai-devkit setup commands listed in the skill and requires telling the user why. Host-config writes are the skill's purpose — kept, but scoped.
  • Also: converts npx ai-devkit@latest fallback to bare npx ai-devkit — prefers an installed binary (zero network when present), fetches only when absent; avoids both @latest bypass and pin staleness.
  • Behavior: instruction text only; no executable code changed.

Test plan

  • Single-file diff: skills/ai-devkit-setup/SKILL.md only
  • Repo pre-push hooks passed

Generated with Devin

Restrict the outside-sandbox clause to the ai-devkit status/setup
commands in this skill, require telling the user why, and pin the npx
fallback to 0.70.1. Instruction text only.
@codeaholicguy
codeaholicguy force-pushed the security/ai-devkit-setup-sandbox-scope branch from 862df18 to 4249def Compare October 14, 2026 19:37
@codeaholicguy
codeaholicguy deleted the security/ai-devkit-setup-sandbox-scope branch October 14, 2026 19:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant