Learn-skills.dev gov_migrate

Run workspace governance migration for an already-running OpenClaw workspace.

install
source · Clone the upstream repo
git clone https://github.com/NeverSight/learn-skills.dev
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/NeverSight/learn-skills.dev "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/skills-md/adamchanadam/openclaw-workspace-governance/gov_migrate" ~/.claude/skills/neversight-learn-skills-dev-gov-migrate && rm -rf "$T"
OpenClaw · Install into ~/.openclaw/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/NeverSight/learn-skills.dev "$T" && mkdir -p ~/.openclaw/skills && cp -r "$T/data/skills-md/adamchanadam/openclaw-workspace-governance/gov_migrate" ~/.openclaw/skills/neversight-learn-skills-dev-gov-migrate && rm -rf "$T"
manifest: data/skills-md/adamchanadam/openclaw-workspace-governance/gov_migrate/SKILL.md
source content

/gov_migrate

Purpose

Execute the migration workflow defined by:

  • prompts/governance/WORKSPACE_GOVERNANCE_MIGRATION.md

Hard contract

  1. If
    _control/GOVERNANCE_BOOTSTRAP.md
    /
    _control/REGRESSION_CHECK.md
    is missing, seed them from canonical payload in
    prompts/governance/OpenClaw_INIT_BOOTSTRAP_WORKSPACE_GOVERNANCE.md
    , then continue migration (do not bounce operator back to manual bootstrap path).
  2. Follow the migration prompt exactly (no skipped gates). 2.1 Before execution, validate migration prompt contract at
    prompts/governance/WORKSPACE_GOVERNANCE_MIGRATION.md
    :
    • Must include anti-precheck clause equivalent to
      Do NOT run canonical equality as a pre-change blocker
      .
    • Must include required order equivalent to
      CHANGE first, then canonical equality at QC
      .
    • If either clause is missing, treat workspace prompt as stale and stop with governance remediation:
      • primary:
        /gov_setup upgrade
        then
        /gov_migrate
      • fallback:
        /skill gov_setup upgrade
        then
        /skill gov_migrate
    • Do not run old pre-change canonical precheck flow.
  3. Preserve non-target user files.
  4. After migration, instruct operator to run
    /gov_audit
    .
  5. Treat workspace root as runtime-resolved
    <workspace-root>
    ; do not hardcode
    ~/.openclaw/workspace
    .
  6. For OpenClaw system claims (commands/config/plugins/skills/hooks), verify using:
    • relevant local skill docs under
      skills/
    • official docs at
      https://docs.openclaw.ai
    • official releases at
      https://github.com/openclaw/openclaw/releases
      for latest/version-sensitive claims
    • if verification cannot be completed, report uncertainty and required next check; do not infer
  7. For date/time-sensitive claims, verify runtime current time context first (session status).
  8. If the operator asks to change platform control-plane state (for example
    ~/.openclaw/openclaw.json
    ), route execution to
    gov_openclaw_json
    and do not patch platform files inside
    gov_migrate
    .
  9. Brain Docs routing:
    • If the task touches Brain Docs (
      USER.md
      ,
      IDENTITY.md
      ,
      TOOLS.md
      ,
      SOUL.md
      ,
      MEMORY.md
      ,
      HEARTBEAT.md
      ,
      memory/*.md
      ), treat read-only asks as Mode B and any write/update as Mode C.
    • For Brain Docs writes, missing READ evidence is fail-closed.
    • For conservative Brain Docs behavior audits/fixes, route to
      gov_brain_audit
      (single entry; preview by default).
  10. Coding-task routing:
  • Any request that creates or modifies workspace code/files (for example: build, implement, fix, refactor) is Mode C, even without
    /gov_*
    command wording.
  • If write intent is uncertain, treat as Mode C (Fail-Closed).
  1. Canonical self-check behavior:
  • If canonical equality check reports mismatch on AUTOGEN core blocks, perform one deterministic repair pass (re-apply canonical AUTOGEN inner content, keep markers unchanged), then re-run equality once before declaring
    BLOCKED
    .
  • Do not treat pre-change mismatch or historical mismatch run reports as immediate blockers.
  • Required sequence: CHANGE first, then canonical equality + optional repair pass at QC.
  1. Official-flow compatibility SOP (anti-self-lock):
  • Treat OpenClaw system operation flows (
    openclaw ...
    , including plugin-added/future commands) and governance lifecycle flows (
    gov_help
    ,
    gov_setup quick/check/install/upgrade
    ,
    gov_migrate
    ,
    gov_audit
    ,
    gov_openclaw_json
    ,
    gov_brain_audit
    ,
    gov_uninstall quick/check/uninstall
    ) as default ALLOW/ROUTE paths, not generic blocks.
  • If a hard prerequisite requires block, label it as governance policy gate (not system error) and provide copy-paste unblock commands.

Output requirements

  • Include
    FILES_READ
    (exact paths) and
    TARGET_FILES_TO_CHANGE
    (exact paths).
  • If either field is missing, output
    BLOCKED (missing read/change evidence)
    and stop.
  • Use this output order for UX consistency:
    1. STATUS
    2. WHY
    3. NEXT STEP (Operator)
    4. COMMAND TO COPY
  • Always include a final
    NEXT STEP (Operator)
    section.
  • If migration PASS:
    • primary:
      /gov_audit
    • fallback:
      /skill gov_audit
  • If migration FAIL or BLOCKED:
    • primary:
      fix blocker, then rerun /gov_migrate
    • fallback:
      fix blocker, then rerun /skill gov_migrate

Fallback

  • If slash command is unavailable or name-collided, use:
    • /skill gov_migrate