Atl live validate
Validate ATL behavior against configured Jira, Confluence, GitHub, or GitLab backends while preserving privacy, request bounds, write authority, owned-fixture scope, and ambiguous-outcome safety. Use for live integration, smoke, compatibility, VPN/connectivity, or fixture checks. Do not use for synthetic unit tests alone.From its SKILL.md
npx -y skills add isukharev/atl --skill atl-live-validateAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.4 KB, as published. Nobody here has run it
Validate ATL live
- Read repository
AGENTS.md; it is binding. - Read Live validation.
- Prove what can be proven synthetically before contacting a backend.
- Load only the isolated configuration identified by current owner instructions. Begin with bounded read-only status/doctor and exact reads.
- Keep all observations generic in public output. Never echo or persist real hosts, IDs, titles, fields, values, content, credentials, or screenshots.
- Before any write or deletion, present the exact owned targets, content, order, request cap, terminal state, reconciliation, and cleanup plan; require explicit authority for that plan.
- Send non-replay-safe writes once. Reconcile any ambiguous outcome before considering another request. Preserve and classify failures privately.
Configured access is not write authority. Stop when a VPN/external dependency must be repaired by its owner or a requested object is not owned by this work.