Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Host security hardening and risk-tolerance guidance for Understudy deployments. Use when a user asks for security audits, firewall/SSH/update hardening, exposure review, or recurring checks on a machine running Understudy.
.claude/skills/understudy-ai-healthcheck/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 20% | 0% |
Assess the host that runs Understudy, keep host controls separate from Understudy configuration, and require explicit approval before any change.
uname -a, sw_vers, cat /etc/os-releasess -ltnup, ufw status, firewall-cmd --state, nft list rulesetlsof -nP -iTCP -sTCP:LISTEN, /usr/libexec/ApplicationFirewall/socketfilterfw --getglobalstate, pfctl -s infotmutil statusunderstudy security --auditunderstudy status --jsonunderstudy health --jsonunderstudy logs --tail 100understudy doctor --deep when broader local diagnostics are neededAsk once for permission to run read-only checks. If granted, infer as much as possible from commands before asking follow-up questions.
Offer one of these numbered defaults after the host context is known:
Always present the plan before making changes. Include:
For each state-changing step:
Require explicit approval for:
After any change, re-check:
understudy security --auditDeliver a final posture summary and note any deferred items.
Use Understudy's canonical scheduling surfaces:
schedule tool with actions status, create, list, update, remove, run, runsunderstudy schedule command with supported flags such as --status, --list, --add, --update, --remove, --run, and --runshealthcheck:security-audithealthcheck:status-snapshotschedule action "list" or understudy schedule --list.understudy security --auditunderstudy status --jsonDo not create recurring jobs without explicit approval.
Use only commands that exist in this repo:
understudy security --auditunderstudy status --jsonunderstudy health --jsonunderstudy logs --tail 100understudy doctor --deepDo not invent understudy update status, understudy cron ..., unsupported schedule subcommands, or flags that are not in this repo.
If the user wants an audit trail, record:
Never log tokens or full credential contents.
Only write to memory files when the user explicitly opts in and the session is private or local. If memory is requested:
memory/YYYY-MM-DD.mdMEMORY.md only for durable preferences such as risk posture or allowed portsIf the session cannot write to the workspace, provide a redacted summary the user can store manually.
Other measured skills in the registry, with their headline benchmark lift.