Config Backup Management
config-backup-management
Capture a point-in-time copy of a network site's configuration as a portable file, track the capture job, retrieve the file, and redeploy a saved copy to recreate or refresh a site (a redeploy, not a documented rollback). Covers Prisma SD-WAN (site snapshots in YAML) today. Use for taking a configuration backup before a change, keeping a baseline, or replicating or refreshing a site from a saved copy. Assumes an authenticated client for the target platform's API is already available; this skill does not provision or authenticate it. Check references/INDEX.md for which platforms and operations are covered.
git clone https://github.com/itential/infra-agent-skills.gitcp -r infra-agent-skills/skills/<skill-name> ~/.claude/skills/Config Backup Management
Captures a saved copy of a configuration, follows the capture to completion, retrieves the copy, and redeploys a saved copy, on whichever platform the caller targets, following that platform’s own documented procedure exactly.
When to use this skill
- Taking a configuration backup or baseline before a change, replicating a site’s configuration, or
refreshing a site by redeploying a previously saved copy. Which of these a platform supports is
platform-dependent; see
references/INDEX.md. - Not for comparing device state before and after a change — see
network-state-collection. - Not for the change record or its approval — see
change-management. - Not for software image upgrades or rollback — see
software-upgrade-lifecycle. - Not for editing the contents of a saved copy or designing configuration; this skill moves copies around, it does not author configuration.
Universal concepts (no wire standard; one implementation today)
Only one platform is documented, so these are expectations that do not extend automatically to a platform not listed in
references/INDEX.md. Where a platform differs, its own procedure wins.
- A saved copy is a file of configuration at a moment in time. Where the platform exposes it, it is produced by a capture request, not by editing live state.
- Capture may be asynchronous. An accepted request is not a finished copy; confirm the job outcome before relying on the file.
- Redeploying a copy writes configuration to the platform. Where a copy matches something that already exists, the effect on that existing object is platform-defined; treat redeploy as a state-changing action and confirm before it.
- A copy can carry sensitive configuration (assumption). Where the platform does not state what is excluded, store and share the file as sensitive.
- Retention, scheduling, and storage location are not defined by the platform procedure unless its own file says so; they are the caller’s concern.
Procedure
- Identify the target platform — from the caller’s request or the
platformargument. Checkreferences/INDEX.md; if no matching row exists, stop: this skill doesn’t cover it yet. - Open
references/<platform>/procedure.mdand follow it exactly — it contains that platform’s full operational procedure, error handling, patterns, known limitations, tools, and verification checklist. - Report the outcome using that platform procedure’s own status vocabulary — don’t normalize it away.
Known limitations
- Platform coverage is in
references/INDEX.md; a platform not listed there needs a new reference file before this skill can be used for it. - Does not cover scheduling, retention, or off-platform storage of copies.
- Does not cover configuration comparison or drift detection.
Contributing to this skill
SKILL.md only changes for two reasons: a bug in the universal concepts/router logic above, or
adding a pointer to a newly-supported platform. Platform-specific procedure changes — a new error
code, a corrected field, a changed API route — belong in that platform’s own
references/<platform>/ files, never here.
Tools
Each platform assumes its own authenticated client — see references/INDEX.md for which, and that
platform’s procedure.md for the specific operations.
Verification checklist
- Target platform confirmed against
references/INDEX.mdbefore proceeding - That platform’s own
procedure.mdwas followed, including its confirmation gate before redeploy - Outcome reported using the platform procedure’s own status vocabulary