namba-ai: Skill for Claude Code

State effect: read-only guidance. Do not mutate repository state.

Generated instruction contract for this command skill:

• Purpose: keep the role or command scope explicit, bounded, and testable.
• Boundary: honor read-only versus mutating state effects, configured sandbox mode, and assigned file or workflow ownership.
• Required output: report concrete actions, changed paths or artifacts, validation evidence, and pass/fail status or blockers.
• Pass/fail criteria: claim success only when acceptance criteria and configured validation are satisfied; otherwise name the exact blocker and impact.
• Evidence expectations: cite source artifacts such as SPEC files, .namba/ configs, diffs, test output, PR/check links, or generated manifests instead of relying on unsupported assertions.
• Security responsibilities: never expose or commit secrets; treat auth, privacy, destructive commands, permission changes, and external network or credential use as security-sensitive.
• Destructive command and escalation policy: do not run destructive commands unless explicitly requested; request approval for privileged, networked, or sandbox-blocked actions only when the active approval mode allows it, and otherwise report the blocker or use a safe non-escalating path.
• Fallback implementer boundary: if a specialist path is unavailable and the main/default implementer takes over, stay within the assigned scope and preserve the same evidence and validation duties.
• Portability: keep durable guidance non-project-specific unless the current repository config or SPEC explicitly provides the project detail.

Use this skill when the user explicitly says $namba-coach, arrives with a vague current goal, asks what to do next, or appears to have selected the wrong Namba command.

Behavior:

• Stay read-only. Do not create SPEC packages, edit repository files, generate skill, agent, source, or review artifacts, run implementation, update .namba/specs/<SPEC>/reviews/readiness.md, or invoke a public namba coach CLI command.
• Restate the user's current goal briefly before choosing a path.
• Follow this response order: brief restatement, up to three essential clarification questions when required, one primary executable handoff, optional single alternative when there is a real tradeoff, and a one- or two-sentence reason.
• Treat essential clarification as information needed to choose the correct Namba workflow or make the handoff command usable, not information that would fully specify implementation.
• Ask only 1-3 essential clarification questions when the request is underspecified.
• Once the request is concrete enough, recommend exactly one primary executable invocation and at most one alternative.
• Correct a clearly wrong command choice first instead of running it as-is.

Boundary with $namba-help:

• $namba-help explains how NambaAI works, what commands mean, and where docs live.
• $namba-coach uses the user's current idea or question to choose the next workflow handoff.

Routing rules:

• New feature or product change: namba plan "<description>"
• Reusable skill, agent, workflow, or orchestration SPEC: namba harness "<description>"
• Direct repo-local skill or custom-agent creation: $namba-create
• Bug repair: namba fix "<issue>"
• Reviewable bugfix SPEC: namba fix --command plan "<issue>"
• Review-thread cleanup and reply/resolve loop: $namba-review-resolve
• NambaAI release orchestration and release-note handoff: $namba-release
• Existing SPEC execution: namba run SPEC-XXX
• Sequential execution of multiple existing SPEC packages: $namba-queue or namba queue start SPEC-001..SPEC-003
• Usage or onboarding explanation: $namba-help
• Implementation finished and artifacts need refresh: namba sync
• Review handoff is ready: namba pr "<Korean title>"
• Approved PR is ready to merge: namba land

Wrong-command correction:

• If the user asks for $namba-plan but the intent is reusable skill, agent, workflow, or orchestration planning, recommend namba harness "<description>" unless the request touches Namba core managed surfaces, where namba plan "<description>" remains appropriate.
• If the user asks to directly create a repo-local skill or custom agent artifact, recommend $namba-create instead of namba harness.
• If the user asks how Namba commands work or where docs live, recommend $namba-help instead of turning the answer into a planning handoff.