agentleFS
vlv-codevlv-code/meta-project/.agents/skills/decommission-meta-project/SKILL.md

decommission-meta-project

Skill0 starsChanged yesterdayMITView on GitHub

Safely decommission, prune, or hibernate agent team roles (.agents/team/): performs non-destructive archiving into .agents/team/archive/, finalizes telemetry, synchronizes root AGENTS.md, and reassigns orphaned tasks without data loss.

  • Deletes or force-pushes

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 1 is distinctive, and 5 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.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before performing any file operations:

    line 28
  2. Very specific:

    Read .agents/meta-project.json and enumerate every binding that references the target role, per installation.

    line 33
  3. Very specific:

    Verify integrity after changes: python scripts/verify_identity.py --check must pass (0 errors) — a dangling profile_path or stale binding …

    line 38
  4. Very specific:

    Ensure target directory .agents/team/archive/ exists.

    line 51
  5. Very specific:

    Remove the decommissioned role's link and description from Section ## 7.

    line 54
  6. Very specific:

    Leave root AGENTS.md and .agents/project_knowledge/ slices intact as the repository's permanent collective memory.

    line 88

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Agent Team Decommissioning and Archival (Meta-Project Decommission)

Autonomous skill for safe, non-destructive decommissioning, resizing, and archiving of specialized agent roles in .agents/team/. Adheres to the core zero-data-loss invariant: physical deletion (rm, Remove-Item) of role workspaces is strictly prohibited. Accumulated context in private notes.md and historical telemetry in profile.md are atomically preserved via archiving into .agents/team/archive/<role_id>/.


When to invoke skill

  • The user requests: "decommission team", "archive agent X", "sunset role X", "hibernate agent team", "prune inactive roles", "prepare team for re-deployment", "сверни команду", "законсервируй агентов", "выведи роль X", "удали агента X из проекта".
  • Project undergoes technology deprecation (e.g., Docker decommissioned, frontend removed, or team resized to Compact Trio).
  • Repository enters long-term maintenance or dormant hibernation mode.

Execution steps

Phase 0: Safety & Scope Verification

Before performing any file operations:

  1. Scan active role directories in .agents/team/ (excluding archive/).
  2. Confirm the target roles to decommission with the user (unless an explicit full reset was commanded).
  3. Core Invariant (Zero Data Loss): Under no circumstances may an agent execute destructive deletion (rm -rf, Remove-Item -Recurse) on role workspaces. Workspaces must always be preserved via atomic relocation to .agents/team/archive/.
  4. Registry-Aware Decommissioning (multi-harness model):
    • Read .agents/meta-project.json and enumerate every binding that references the target role, per installation.
    • Distinguish two different operations:
      • (A) Disable ONE installation's binding (e.g. disconnecting ZCode while Antigravity keeps working): set that binding's status to "disabled" in the registry and archive ONLY the installation-specific profile copy (e.g. .agents/team/<role_id>-<installation_suffix>/) if one exists. The shared role, other bindings, other installations' profiles, notes, and telemetry stay untouched.
      • (B) Archive the SHARED role: allowed ONLY when NO other active or planned installation still references the role. If any other installation references it, REFUSE with an explicit report listing the referencing installations and offer operation (A) instead.
    • Apply registry changes, the AGENTS.md Section 7 roster, and the archive relocation CONSISTENTLY in one pass; preserve memory and historical authorship (profile.md telemetry, notes.md) byte-for-byte.
    • Verify integrity after changes: python scripts/verify_identity.py --check must pass (0 errors) — a dangling profile_path or stale binding is a failed decommission.

Scenario A: Targeted Role Sunsetting (Single Role Pruning)

Used when a single technology layer or specialization is retired from the project.

  1. Telemetry Closeout: Append a closing entry to .agents/team/<role_id>/profile.md in the telemetry table:
    | <YYYY-MM-DD> | User | Decommissioning / Sunsetting | Role archived into .agents/team/archive/<role_id>/ |
  2. Directory Archival:
    • Ensure target directory .agents/team/archive/ exists.
    • Atomically move .agents/team/<role_id>/ to .agents/team/archive/<role_id>/.
  3. Passport Synchronization (AGENTS.md):
    • Remove the decommissioned role's link and description from Section ## 7. Project Agent Team & Multi-Harness Identity (.agents/team/) in root AGENTS.md.
  4. Orphaned Tasks Reassignment:
    • Scan .agents/project_knowledge/tasks.md.
    • Reassign any pending subtasks delegated to this role to the default generalist (agent-general-worker / agent-dev-lead) or mark them as [CANCELLED].
  5. Observation Logging:
    • Record the role sunsetting fact into .agents/project_knowledge/observations.md via record-observation.

Scenario B: Team Resizing & Re-deployment Reset (Batch Archival)

Used when transitioning from a bloated team to a Scale-Adaptive Compact Trio, or rotating team structure.

  1. Batch Telemetry Closeout: For every active role in .agents/team/, append a transition entry in profile.md:
    | <YYYY-MM-DD> | User | Decommissioning / Resizing | Archived during team re-deployment and sizing calibration |
  2. Batch Archival Relocation:
    • Ensure .agents/team/archive/ exists.
    • Move all active role directories from .agents/team/<role_id>/ into .agents/team/archive/<role_id>/. (If a role with the same ID already exists in archive, append timestamp suffix: .agents/team/archive/<role_id>-<YYYYMMDD>/).
  3. Passport Section Reset:
    • In root AGENTS.md, clear the decommissioned entries under ## 6. Project Agent Team (.agents/team/) to prepare for the new synthesized team.
  4. Tasks Quarantine:
    • In .agents/project_knowledge/tasks.md, close or quarantine active tasks from the prior team cycle.
  5. Report & Handoff:
    • Report archival completion to the user. The repository is now clean and ready for manage-meta-project to materialize the new synthesized team.

Scenario C: Full Repository Hibernation (Dormant Mode)

Used when active development concludes and the repository transitions into a dormant maintenance state.

  1. Execute Scenario B steps to archive all active roles into .agents/team/archive/.
  2. Leave root AGENTS.md and .agents/project_knowledge/ slices intact as the repository's permanent collective memory.
  3. Inform the user: "All agent roles have been safely hibernated into .agents/team/archive/. Static project memory and knowledge slices remain active. Re-activate anytime via skills/manage-meta-project/SKILL.md."

Phase Final: Decommissioning Report

Upon completing any scenario, output a concise status table:

| Decommissioned Role | Archive Destination Path | Unfinished Tasks Handled | Status |
| :--- | :--- | :--- | :--- |
| `agent-deep-auditor` | `.agents/team/archive/agent-deep-auditor/` | Reassigned / None | Archived |
| `agent-qa-engineer` | `.agents/team/archive/agent-qa-engineer/` | Reassigned / None | Archived |

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.