proxmox/docs/00-meta/ROOT_DIRTY_WORK_REVIEW_2026-03-29.md

Root Dirty Work Review — 2026-03-29

Purpose: Separate unrelated root-repo dirty work from the DBIS RTGS tranche so operators can review risk and merge intent without conflating changes.

Scope reviewed

From git status --short in the root repo:

• modified:
  • docs/04-configuration/GITEA_ACT_RUNNER_SETUP.md
  • pnpm-lock.yaml
  • scripts/dev-vm/setup-act-runner.sh
  • scripts/lib/load-project-env.sh
  • smom-dbis-138 submodule pointer dirty
• untracked:
  • scripts/dev-vm/bootstrap-gitea-act-runner-site-wide.sh
  • scripts/dev-vm/install-act-runner-systemd.sh

Findings

1. Gitea act-runner lane

Files

• GITEA_ACT_RUNNER_SETUP.md
• setup-act-runner.sh
• bootstrap-gitea-act-runner-site-wide.sh
• install-act-runner-systemd.sh

What changed

• docs now describe a site-wide bootstrap path
• setup defaults now target http://127.0.0.1:3000
• runner labels default to ubuntu-latest
• new scripts appear intended to:
  • fetch a registration token via admin API
  • install and daemonize the runner under systemd

Assessment

• This is a coherent feature lane, not random drift.
• It should be reviewed and committed as a dedicated act_runner bootstrap tranche.
• It is unrelated to the RTGS sidecar work.

Risk

• low to medium if kept isolated
• medium to high if accidentally bundled with RTGS or infra-truth updates

2. scripts/lib/load-project-env.sh

What changed

• VMID 5700 was added to the r630-02 host mapping set.

Assessment

• This appears to be a small support fix for the act-runner / dev-vm lane.
• It is logically related to the act-runner work above.
• It should ship with that lane, not with unrelated RTGS work.

Risk

• low if correct
• moderate if 5700 is mobile and the mapping is treated as permanently canonical

3. pnpm-lock.yaml

What changed

• substantial additions for a root importer named smom-dbis-138
• many Hardhat / TypeChain / changesets and related packages

Assessment

• this is a generated artifact from a package-manager operation
• it likely belongs to a separate smom-dbis-138 JavaScript / Hardhat tooling lane
• it is not tied to the RTGS docs/deployment tranche

Risk

• medium if committed without the matching package.json / workspace intent
• high if it accidentally lands in the root repo without the team intending to manage smom-dbis-138 from the root PNPM workspace

4. smom-dbis-138 submodule dirty state

Observed

• deleted contract files:
  • contracts/emoney/BridgeVault138.sol
  • contracts/emoney/ComplianceRegistry.sol
  • contracts/emoney/PolicyManager.sol
  • contracts/emoney/TokenFactory138.sol
  • several contracts/emoney/interfaces/*
• modified:
  • package.json
  • package-lock.json
• untracked:
  • playwright-report/

Assessment

• this is high-risk and should be treated as a separate active refactor or deletion lane
• the deleted contracts are core-emoney-surface files, not cosmetic churn
• do not bundle or auto-commit this with unrelated infra/docs work

Risk

• high

Recommended split

1. Act-runner tranche

   • docs
   • setup script
   • bootstrap script
   • systemd install script
   • load-project-env.sh mapping

2. Root JS / lockfile tranche

   • pnpm-lock.yaml
   • only if intentionally paired with a matching workspace/package change

3. smom-dbis-138 contract/package tranche

   • separate review required
   • verify whether the deletions are intentional refactor, move, or accidental loss

Rule

Until these lanes are reviewed and intentionally grouped, they should remain excluded from RTGS-sidecar, explorer, and DBIS master-plan commits.