sdd-continue

Follow the SDD orchestrator workflow to continue the active change.

HARD GATE: SDD Session Preflight must already be complete for this session. It must include execution mode, artifact store, chained PR strategy, and review budget. If missing, ask the exact orchestrator preflight prompt and STOP. Do not launch the next phase in the same turn.

WORKFLOW:

1. If the gentle-ai binary is available, run gentle-ai sdd-continue [change] --cwd <repo> and treat its dispatcher/status output as authoritative — but only when the session artifact store is openspec or hybrid. When the session artifact store is engram, do NOT invoke the native dispatcher at all — it cannot see the change (it reads only openspec/changes/); resolve status entirely from Engram (mem_search + mem_get_observation on the change's topic keys) using the manual status schema in ~/.config/opencode/skills/_shared/sdd-status-contract.md (the same schema used when the binary is unavailable). The dispatcher is authoritative only for openspec/hybrid. If unavailable, resolve the active change using the status contract. If $ARGUMENTS is missing and more than one active change exists, ask the user to choose and STOP. Do not guess.
2. Produce or consume structured status before acting: schemaName, planningHome/changeRoot, artifactPaths/contextFiles, task progress, dependency states, next recommended action, blocked reasons, and actionContext.
3. Check which artifacts already exist for the active change (proposal, specs, design, tasks)
4. Determine the next phase needed based on the dependency graph: proposal → [specs ∥ design] → tasks → apply → verify → archive
5. Launch the appropriate sub-agent(s) for the next phase only if authoritative status says the dependency is ready. Route only by nextRecommended and dependency states; never infer from free text. If blockedReasons is non-empty, do not proceed to apply, archive, or terminal work. If nextRecommended is verify, verification/remediation may run only to refresh evidence; if nextRecommended is resolve-blockers, report blockedReasons and stop; if nextRecommended is a planning token (propose, spec, design, or tasks), launch the corresponding planning phase.
6. Present the result and ask the user to proceed

CONTEXT:

• Working directory: before doing anything else, run git rev-parse --show-toplevel 2>/dev/null || pwd with your bash tool and use the returned path as the authoritative workspace. In OpenCode Desktop (Electron) the parse-time interpolation resolves to the app data directory, not the project.
• Current project: the basename of the detected workspace above.
• Change name: $ARGUMENTS
• Execution mode: ask/cache per orchestrator
• Artifact store mode: ask/cache per orchestrator; do not hardcode Engram
• Delivery strategy: ask/cache per orchestrator
• Review budget: ask/cache per orchestrator

ENGRAM NOTE: To check which artifacts exist in engram/hybrid, search: mem_search(query: "sdd/$ARGUMENTS/", project: "{project}") to list all artifacts for this change. Sub-agents handle persistence automatically using the selected artifact store.

Read the orchestrator instructions to coordinate this workflow. Do NOT execute phase work inline — delegate to sub-agents.

STATUS CONTRACT:

Prefer gentle-ai sdd-continue [change] --cwd <repo> when available — but only when the session artifact store is openspec or hybrid; when the store is engram, do NOT invoke the binary and resolve status from Engram using the manual status schema. Otherwise read the installed shared status contract from this agent's skills directory and follow it. Use ~/.config/opencode/skills/_shared/sdd-status-contract.md for OpenCode, ~/.config/kilo/skills/_shared/sdd-status-contract.md for Kilo Code, ~/.qwen/skills/_shared/sdd-status-contract.md for Qwen, or the equivalent configured skills directory for the current adapter. Do not use a workspace-relative skills/_shared/... path. Carry actionContext and allowed edit roots into any sub-agent launch. If status reports workspace-planning with no allowed edit roots, do not launch apply/verify/archive work that would infer repo-local ownership.