gentle-sdd-archive

If the native sdd-archive sub-agent is available, delegate this command to it. Otherwise, read the skill file at ~/.claude/skills/sdd-archive/SKILL.md FIRST, then follow its instructions exactly inline.

CONTEXT:

• Working directory: Detect agent-side before proceeding by running git rev-parse --show-toplevel with the Bash tool; if that fails, run pwd with the Bash tool.
• Current project: Derive agent-side from the detected working directory basename. Do not use slash-command shell interpolation for this value.
• Artifact store mode: engram

TASK: Archive the active SDD change. Read the verification report first to confirm the change is ready. Then:

STATUS GATE: Read ~/.claude/skills/_shared/sdd-status-contract.md and produce structured status before acting. If $ARGUMENTS is missing or ambiguous, ask the user to choose and STOP. Do not guess. Archive only when refreshed native SDD status reports dependencies.archive: ready and nextRecommended: archive. Continue only when verify-report exists, contains no CRITICAL issues, and tasks are complete. reviewOffer is optional and never an archive or delivery gate. Archive reads only task completion and independent verification; a review offer may be declined or completed without changing archive readiness. CRITICAL verification issues have no override. If unchecked tasks remain, send the change back to sdd-apply unless apply-progress/verify-report prove they are stale checkboxes and the orchestrator explicitly requests mechanical reconciliation. If status reports workspace-planning, STOP and explain that workspace archive is not supported in this slice. Carry contextFiles, task progress, dependency states, and actionContext into the native sub-agent prompt when delegating.

ENGRAM PERSISTENCE (artifact store mode: engram): CRITICAL: mem_search returns 300-char PREVIEWS, not full content. You MUST call mem_get_observation(id) for EVERY artifact. STEP A — SEARCH (get IDs only): mem_search(query: "sdd/{change-name}/proposal", project: "{project}") → save proposal_id mem_search(query: "sdd/{change-name}/spec", project: "{project}") → save spec_id mem_search(query: "sdd/{change-name}/design", project: "{project}") → save design_id mem_search(query: "sdd/{change-name}/tasks", project: "{project}") → save tasks_id mem_search(query: "sdd/{change-name}/verify-report", project: "{project}") → save verify_id STEP B — RETRIEVE FULL CONTENT (mandatory): mem_get_observation(id: proposal_id) → full proposal mem_get_observation(id: spec_id) → full spec mem_get_observation(id: design_id) → full design mem_get_observation(id: tasks_id) → full tasks mem_get_observation(id: verify_id) → full verification report Record all observation IDs in the archive report for traceability. Treat verify-report and apply-progress as intermediate snapshots: the archive report records the state at close per the skill's Final-State Authority section, and explicit final-state facts in your launch prompt outrank stale snapshot claims.

Archival is a MECHANICAL filesystem operation: copy/move artifacts with cp -R/mv/git mv via the shell only — NEVER via model Read/Write, which can truncate bytes silently while reporting success. After every copy/move, run diff -r (source vs. destination, archive-report additive-only) and include its verbatim output in the result; an empty diff is the only passing evidence, and a skipped/missing diff -r FAILS the phase. If shell access is unavailable, STOP and report blocked — do NOT fall back to Read/Write copying. Save: mem_save(title: "sdd/{change-name}/archive-report", topic_key: "sdd/{change-name}/archive-report", type: "architecture", project: "{project}", capture_prompt: false, content: "{archive report with observation IDs}") Set capture_prompt: false when the Engram tool schema supports it; if an older schema rejects or does not expose the field, omit it rather than failing.

Then:

1. Sync delta specs into main specs (source of truth)
2. Move the change folder to archive with date prefix
3. Verify the archive is complete