dshcode: Skill for Claude Code

mancode mode: manteam

Purpose: plan and execute work with explicit team ownership and handoff.

Enter through mancode

Before the first command, use ./node_modules/.bin/mancode when it exists, otherwise mancode; check that selected binary with --version once and never mix binaries or versions. In every command below, replace the literal mancode with that selected binary path when the local binary exists.

1. Reuse a mancode status --brief --json snapshot already obtained in this conversation. Only when none exists, run it once from the project root. Never read or write the legacy authority file.
2. If identity.actorId is absent, ask for a display name and run mancode team identity create --name "<display name>".
3. Reuse session.sessionId when present. If status has no current session, reuse an explicit session ID already retained in this conversation. Only if neither exists, run mancode context session new --client codex in Codex, mancode context session new --client zcode in ZCode, or mancode context session new --client kimi-code in Kimi Code exactly once, then retain the returned session ID. Use --client codex in Codex, --client zcode in ZCode, or --client kimi-code in Kimi Code on every command that uses this session.
4. Reuse the current TaskRef. To bind a supplied existing task, run mancode context resume <namespace:ULID> --session <id> --client <active-client>.
5. For an existing task, read only the needed Context Pack with mancode context show --purpose plan --session <id> --client <active-client>; include --task <namespace:ULID> when it is not yet bound. For a new task, create it through the mode action first, then read the returned TaskRef's Context Pack.

Mode action

• Confirm team membership with mancode team status; join invited participants before assigning shared work.
• Read .mancode/shared/context/glossary.json when it exists and prefer its confirmed terms in shared requirements, plans, and handoffs.
• For a new shared task, run mancode workflow create manteam "<task>" --visibility shared --coordination team --confirm-shared --session <id> --client <active-client>.
• Run the same bounded read-only discovery as man, with at most F-1 through F-3 typed as premise, scope, technical, risk, or acceptance and marked as repository_fact or domain_hypothesis; an unverified domain hypothesis becomes a focused question, never a fact. Discovery produces evidence and recommendations, never execution authority.
• Give every F-ID one type-directed disposition: Accepted scope or behavior findings enter confirmedScope and the matching acceptanceCriteria; accepted technical choices enter technicalDecisions; Only explicitly excluded behavior enters excludedScope; low-impact reversible details may enter defaults; an unaccepted proposal remains unauthorized without being copied into every field.
• Apply the same decision-readiness gate as man before finalizing requirements: validate both clarity and soundness against project facts and team authority. If the goal, scope, acceptance, owner/source of truth, and constraints are clear and consistent, continue without ceremonial questions; if a decision-changing ambiguity, ownership conflict, or hard-risk direction remains, give evidence and a recommendation, ask focused questions, and wait before writing confirmed requirements.
• Persist unresolved team clarification through the same workflow requirements <namespace:ULID> draft --file <requirements.json> command as man; do not leave ownership questions or partial answers only in chat history.
• Bind the confirmed team plan to the user-visible implementationScope through plan revise --scope-file; claims, edits, and review must stay inside include and outside exclude. Before editing, state material assumptions and verifiable success criteria, reuse existing code and dependencies, and make the smallest direct plan-traceable change; newly proposed behavior outside confirmed requirements requires read-only NEEDS_REALIGNMENT and operator-approved reframe.
• If the operator explicitly approves a file-boundary-only adjustment that leaves confirmed behavior and acceptance unchanged, use mancode workflow scope change <shared:ULID> --expected-revision <n> --file <scope.json> --session <id> --client <active-client>. It versions the plan authority, stales prior review/verification, and reissues compatible claims. Behavior or acceptance changes still require reframe.
• Use claims, checkpoints, sync, and handoffs through mancode team; never infer ownership from an adapter prompt.
• With git-ref transport, workflow creation plus requirements, plan, review, and verification mutations use an explicit deferred publication boundary: run the workflow command without --sync, commit the resulting .mancode/shared authority changes together with the matching code head, then run mancode team sync push <shared:ULID> --expected-task-revision <n>. Never report cross-clone synchronization before the push returns a receipt. Use --sync only for a command whose contract performs an atomic git-ref mutation. If that atomic mutation leaves tracked .mancode/shared projection changes for a resumable in-progress or blocked task, commit them, then run the same team sync push with the unchanged task revision to rebind the remote code head before another clone resumes the task.