agentleFS
ghivelyghively/infra-ops/skills/perso-change-control/SKILL.md

perso-change-control

Skill0 starsChanged 5 months agoMITView on GitHub

In-zone (HSA) change-control process for personalization/data-prep systems: CISO-approved, dual-control, witnessed test-to-live promotion with separation of duties between dev and production staff. The agent prepares change artifacts and records; humans approve and deploy. Advisory-only; never an autonomous deploy. Triggers on: change control, test to live, promotion, dual control, witnessed sign-off, ciso approval, separation of duties, perso, personalization, in-zone, hsa, cp logical 6.2, emergency change, baseline validation.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Test → live promotion is a human, witnessed, dual-control event:

    line 34

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Perso Change-Control Skill

ADVISORY STUB — Phase 7 / CPSA-gated. Not yet active.

Encodes the in-zone change-control discipline from PCI Card Production Logical §6.2–§6.6. The agent's role is to prepare compliant change artifacts and gate on the controls — never to approve or deploy. Verify section numbers against the current PCI CP Logical PDF. Do not extend without a CPSA review.

When to Use

Load this skill whenever a perso-* agent prepares or reviews a change destined for the personalization or data-preparation network: any promotion from test toward live, any emergency change, or any baseline-validation activity. It defines the human gates the agent must build toward and must never bypass.

How It Works

The Promotion Gate (CP Logical §6.2)

Test → live promotion is a human, witnessed, dual-control event:

  1. CISO approval — documented before deployment. No change reaches live without it.
  2. Dual sign-off — both development and production staff sign off, witnessed under dual control. Two distinct people; the agent is never one of them.
  3. Separation of duties (§6.6.3 d) — the identity that authored the change cannot be the identity that deploys it to live perso. Enforce author ≠ deployer.
  4. No autonomous deploy — the agent opens an in-zone MR and prepares the change record; a human performs the promotion. The dual-control-promotion-gate is the machine check (two distinct approvers + citation + CPSA reference + in-zone flag).

The Change Record the Agent Prepares

For every in-zone change, the agent assembles (text/artifacts only — no PAN/keys/PINs):

  • What & why — scope, target systems, and the business/security justification.
  • --check --diff evidence — a dry-run diff proving the change is idempotent.
  • Rollback plan — explicit, tested revert path (per rollback-and-runbooks).
  • Blast radius — which perso/DP hosts are affected; what breaks if it fails.
  • Approver slots — named dev signer, named prod signer, witness, CISO approver (left for humans to sign; the agent never self-signs).
  • Ceremony notes — if the change is adjacent to a key/PIN ceremony, reference the ceremony and its custodians; never include key/PIN material.

Emergency Change (CP Logical §6.2)

  • A documented emergency procedure with retrospective approval — not a bypass of control, a different, logged path. The agent records the emergency justification and flags it for retrospective CISO sign-off. The INFRAOPS_BYPASS_DUAL_CONTROL flag on the promotion gate is audited and must map to a documented emergency procedure.

Baseline & Patch Discipline (CP Logical §6.3)

  • Monthly configuration validation against the authorized baseline; a non-empty drift is a finding (pairs with drift-detection).
  • Security patches within 30 days (critical / Internet-facing: 7 business days).
  • The agent can propose the patch/validation change and the evidence; humans apply.

Separation of Duties & Source Access (CP Logical §6.6.3)

  • Personalization-application source is restricted to authorized personnel; the in-zone GitLab is separate from corporate GitLab with its own ACLs.
  • Dev and production staff are distinct roles. The agent's service account is least-privilege: branch/MR write only — no protected-branch, no deploy rights.

What the Agent Never Does

NEVER:
  - approve a change (CISO / dev / prod sign-off are human)
  - deploy to live perso/DP systems
  - act as the second person in dual control
  - merge its own in-zone MR
  - include PAN / keys / key components / PINs in any change artifact
ALWAYS:
  - prepare the compliant change record + check/diff evidence + rollback
  - leave every approval slot for a named human
  - route key/PIN ceremonies to the custodian team by reference

Examples

# Correct response when asked to "push this straight to the perso boxes":
"No autonomous deploy to live personalization systems. Test→live requires CISO
approval and a witnessed, dual-control sign-off by both dev and production staff
(CP Logical §6.2). I've prepared the change record, --check --diff evidence, and a
rollback plan; the named approvers must sign and a human performs the promotion."

# Correct response on an emergency change:
"Logged as an emergency change with justification attached for retrospective CISO
approval (CP Logical §6.2). If the dual-control bypass is used it is audited to the
off-box ledger and must reference the documented emergency procedure."

TODO: Flesh out once the in-zone GitLab/Octopus topology and the real perso deployment process are confirmed and ingested, and a CPSA-L assessor has reviewed the design. (DESIGN §14 Phase 7)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.