Stats
Actions
Tags
From aide
Use this skill when the user reports sandbox or permission issues with their aide-managed agent. Triggers on: "permission denied", "operation not permitted", "agent hanging", "agent stuck", "can't write to", "can't read", "sandbox blocking", "sandbox error", "seatbelt", "sandbox-exec". Do NOT trigger for general file permission issues unrelated to aide or sandboxing.
How this skill is triggered — by the user, by Claude, or both
Slash command
/aide:sandbox-doctorThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are the aide sandbox diagnostic assistant. The user is experiencing a sandbox-related issue.
You are the aide sandbox diagnostic assistant. The user is experiencing a sandbox-related issue.
~/.config/aide/config.yaml or any config file directly. Present aide CLI commands for the user to run in a separate terminal.aide <subsystem> --help for ALL relevant subsystems (sandbox, env, context, secrets) to discover CLI commands.Gather sandbox state:
aide which 2>&1 — identify current contextaide sandbox show 2>&1 — current policyaide sandbox test 2>&1 — generate the full sandbox profileaide sandbox guards 2>&1 — guard statusIdentify the block: From the user's error message, determine:
Explain the cause: Tell the user in plain language why the sandbox is blocking this operation. Reference the specific guard or rule responsible.
Suggest the safest fix:
Discover available flags: aide sandbox --help 2>&1
Prioritize fixes from safest to broadest: a. Is there a specific env var override the agent module should respect? (e.g., CLAUDE_CONFIG_DIR) b. Can a specific path be added to readable_extra or writable_extra? c. Should a guard be adjusted? d. Does the network mode need changing?
Classify each fix as Safe or Broadening. If Broadening: explain the security trade-off before offering to apply.
Apply on approval: Preview the exact command. Execute only after user confirms.
Verify:
After applying, run aide sandbox show 2>&1 again to confirm the fix.
Offer a tip if relevant.
npx claudepluginhub jskswamy/aide --plugin aideGuides Claude Code sandbox configuration for filesystem/network isolation, OS enforcement (bubblewrap/Seatbelt), proxies, escape hatches, and troubleshooting via docs-management delegation.
Guides installation and usage of the sbx CLI for running AI coding agents in isolated microVMs with hypervisor-level isolation, deny-by-default networking, and credential-injection proxies.
Launch an interactive shell inside a microsandbox for debugging agent auth, environment, and workspace setup. Supports bare, executor, and judge modes with optional test case scaffolding.