free-jt7-global-runtime-audit

Free JT7 Global Runtime Audit

Overview

Use this skill to run a deterministic audit and hardening flow for Free JT7 runtime behavior. Focus on executable evidence, not assumptions.

Workflow

1. Validate core runtime policy and health.
2. Validate local execution capabilities (filesystem, processes, network).
3. Validate cross-IDE resolution and bridge files.
4. Validate traceability artifacts and checklist status.
5. Apply fixes, rerun checks, and produce evidence paths.

Mandatory Commands

Run these commands in order:

python skills_manager.py policy-validate
python skills_manager.py doctor --strict
python skills_manager.py rollout-mode
python skills_manager.py host-mode status
python skills_manager.py ide-detect --json

Then execute one audited run:

python skills_manager.py task-run --goal "runtime-audit" --commands "Get-ChildItem" "python --version"
python skills_manager.py task-list --limit 10
python skills_manager.py task-checklist --run-id <run_id>

Fix Rules

• If task-list or task-checklist is missing, implement those commands before closing.
• If task runs are not reflected in copilot-agent/tasks.yaml, implement upsert/update logic.
• If Codex user settings are not updated by install --update-user-settings, add support in settings updater.
• If docs contain hardcoded old paths, replace with relative runtime commands.

Evidence

Always provide:

• Latest run JSON path in copilot-agent/runs/.
• Latest checklist output for the validated run.
• Files changed to fix missing behavior.

Reference

For a full command matrix and expected outputs, read: references/runtime-checklist.md