Claude Skill

shukka

Prepare software for responsible release, migration, deployment, rollback, and handoff with evidence about compatibility, health, ownership, and recovery.

LLM Mart · 0 points · 0 views 0 listing impressions 0 install-command copies
Virus-scanned Reviewed automatically before listing.

Full trust report

Download kuya-egg-monozukuri-shukka-b4243fe.zip · 2 KB
Part of kuya-egg/monozukuri — 13 skills

Install

skills CLI npx skills add https://github.com/kuya-egg/Monozukuri/tree/main/shukka
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install kuya-egg-monozukuri@llmmart
Git git clone https://github.com/kuya-egg/Monozukuri.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole kuya-egg/monozukuri collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Shukka

Shukka means shipping or dispatching a finished thing. In software, release is part of the product: move verified behavior into use with operational discipline, clear ownership, and a recovery path. This is a Japanese-inspired engineering metaphor, not a rule that every change needs a ceremony.

Use this skill when

  • preparing a release, deployment, migration, or handoff;
  • changing configuration, infrastructure, schemas, or public contracts;
  • deciding whether a feature is ready for users;
  • staging a high-risk or irreversible change; or
  • defining what must be observed after release.

Outcome

Produce a release decision and handoff containing:

  • exact scope and version or revision;
  • compatibility and migration considerations;
  • required configuration, secrets, permissions, and owners;
  • checks run and their results;
  • health, observability, and success signals;
  • rollout, rollback, and recovery paths;
  • user or operator communication; and
  • residual risk and explicit release blockers.

Workflow

1. Define the release boundary

Identify what changes, who is affected, what must remain compatible, and which data or external state can be changed. Separate reversible from irreversible steps.

2. Verify the artifact

Check the relevant tests, type checks, builds, migrations, contracts, security checks, smoke checks, and operational prerequisites. Do not use “merged” or “compiles” as a release decision.

3. Prepare safe movement

Choose a staged rollout, feature flag, backup, transaction, maintenance window, or rollback mechanism when risk justifies it. Confirm that rollback is real, tested enough to trust, and does not destroy newer data.

4. Prepare observation and ownership

Name the signals that indicate success, degradation, and recovery. Identify who responds, how to find diagnostics, and when to stop or reverse the rollout.

5. Communicate the handoff

State what changed, how to use it, how to operate it, known limitations, and what to do if it fails. Keep the message useful to the next person, not merely ceremonial.

6. Verify after movement

Run smoke or health checks and inspect real signals after release. Record the result and any follow-up. If a meaningful check fails, use jidoka and do not hide the failure behind a successful deployment status.

Evidence standard

Every release claim must name the check, environment, revision, and result when that information matters. If rollback, recovery, or production behavior was not verified, state the limitation explicitly.

Boundaries

  • Do not deploy or mutate external systems without the user's authorization.
  • Do not treat a checklist as evidence when the underlying check did not run.
  • Do not add release complexity without a risk or recovery need.
  • Do not call a migration safe without considering partial failure and rollback.
  • Do not hand off undocumented operational responsibility.

Handoff

End with a release decision: ready, blocked, or ready with explicitly accepted risk. Link the evidence, owners, recovery path, and next reflection. Use andon for missing signals, jidoka for failed checks, and hansei after the release to preserve lessons.

If a referenced skill is not installed, apply its named lens inline instead of trying to invoke it.

Files (monozukuri)
  • agents
    • openai.yaml 329 B
      interface:
        display_name: "Shukka — Responsible Release"
        short_description: "Release changes with recovery discipline"
        brand_color: "#1F2937"
        default_prompt: "Use $shukka to prepare this change for release with compatibility, observability, rollback, and ownership evidence."
      
      policy:
        allow_implicit_invocation: false
      
  • SKILL.md 3.4 KB
    ---
    name: shukka
    description: "Prepare software for responsible release, migration, deployment, rollback, and handoff with evidence about compatibility, health, ownership, and recovery."
    ---
    
    # Shukka
    
    Shukka means shipping or dispatching a finished thing. In software, release is
    part of the product: move verified behavior into use with operational
    discipline, clear ownership, and a recovery path. This is a Japanese-inspired
    engineering metaphor, not a rule that every change needs a ceremony.
    
    ## Use this skill when
    
    - preparing a release, deployment, migration, or handoff;
    - changing configuration, infrastructure, schemas, or public contracts;
    - deciding whether a feature is ready for users;
    - staging a high-risk or irreversible change; or
    - defining what must be observed after release.
    
    ## Outcome
    
    Produce a release decision and handoff containing:
    
    - exact scope and version or revision;
    - compatibility and migration considerations;
    - required configuration, secrets, permissions, and owners;
    - checks run and their results;
    - health, observability, and success signals;
    - rollout, rollback, and recovery paths;
    - user or operator communication; and
    - residual risk and explicit release blockers.
    
    ## Workflow
    
    ### 1. Define the release boundary
    
    Identify what changes, who is affected, what must remain compatible, and which
    data or external state can be changed. Separate reversible from irreversible
    steps.
    
    ### 2. Verify the artifact
    
    Check the relevant tests, type checks, builds, migrations, contracts, security
    checks, smoke checks, and operational prerequisites. Do not use “merged” or
    “compiles” as a release decision.
    
    ### 3. Prepare safe movement
    
    Choose a staged rollout, feature flag, backup, transaction, maintenance window,
    or rollback mechanism when risk justifies it. Confirm that rollback is real,
    tested enough to trust, and does not destroy newer data.
    
    ### 4. Prepare observation and ownership
    
    Name the signals that indicate success, degradation, and recovery. Identify who
    responds, how to find diagnostics, and when to stop or reverse the rollout.
    
    ### 5. Communicate the handoff
    
    State what changed, how to use it, how to operate it, known limitations, and
    what to do if it fails. Keep the message useful to the next person, not merely
    ceremonial.
    
    ### 6. Verify after movement
    
    Run smoke or health checks and inspect real signals after release. Record the
    result and any follow-up. If a meaningful check fails, use `jidoka` and do not
    hide the failure behind a successful deployment status.
    
    ## Evidence standard
    
    Every release claim must name the check, environment, revision, and result
    when that information matters. If rollback, recovery, or production behavior
    was not verified, state the limitation explicitly.
    
    ## Boundaries
    
    - Do not deploy or mutate external systems without the user's authorization.
    - Do not treat a checklist as evidence when the underlying check did not run.
    - Do not add release complexity without a risk or recovery need.
    - Do not call a migration safe without considering partial failure and rollback.
    - Do not hand off undocumented operational responsibility.
    
    ## Handoff
    
    End with a release decision: ready, blocked, or ready with explicitly accepted
    risk. Link the evidence, owners, recovery path, and next reflection. Use
    `andon` for missing signals, `jidoka` for failed checks, and `hansei` after the
    release to preserve lessons.
    
    If a referenced skill is not installed, apply its named lens inline instead of
    trying to invoke it.
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.

Related