skill-status-sync（技能状态同步）

Status Sync Skill (Direct Execution)

Direct execution skill for atomic status synchronization across TODO.md and state.json. This skill executes inline without spawning a subagent, avoiding memory issues.

Context References

Reference (do not load eagerly):

• Path: .claude/context/patterns/jq-escaping-workarounds.md - jq escaping patterns (Issue #1132)

Standalone Use Only

IMPORTANT: This skill is for STANDALONE USE ONLY.

Workflow skills (skill-researcher, skill-planner, skill-implementer, etc.) now handle their own preflight/postflight status updates inline. This eliminates the multi-skill halt boundary problem where Claude may pause between skill invocations.

Use this skill for:

• Manual task status corrections
• Standalone scripts that need to update task state
• Recovery operations when workflow skills fail
• Testing status update behavior in isolation

Do NOT use this skill in workflow commands (/research, /plan, /implement, /revise) - those commands now invoke a single skill that handles its own status updates.

Trigger Conditions

This skill activates when:

• Manual status correction is needed
• Artifacts need to be linked outside normal workflow
• Status synchronization recovery is needed between TODO.md and state.json

API Operations

This skill exposes three primary operations:

Execution

1. Input Validation

Validate required inputs based on operation type:

For preflight_update:

• task_number - Must be provided and exist in state.json
• target_status - Must be an in-progress variant (researching, planning, implementing)
• session_id - Must be provided for traceability

For postflight_update:

• task_number - Must be provided and exist
• target_status - Must be a final variant (researched, planned, implemented, partial)
• artifacts - Array of {path, type} to link
• session_id - Must be provided

For artifact_link:

• task_number - Must be provided and exist
• artifact_path - Relative path to artifact
• artifact_type - One of: research, plan, summary

2. Execute Operation Directly

Route to appropriate operation and execute using Bash (jq) and Edit tools.

Operation: preflight_update

Purpose: Set in-progress status at GATE IN checkpoint

Execution:

1. Validate task exists:

task_data=$(jq -r --arg num "{task_number}" \
  '.active_projects[] | select(.project_number == ($num | tonumber))' \
  specs/state.json)

if [ -z "$task_data" ]; then
  echo "Error: Task {task_number} not found"
  exit 1
fi

1. Update state.json:

jq --arg ts "$(date -u +%Y-%m-%dT%H:%M:%SZ)" --arg status "{target_status}" \
  '(.active_projects[] | select(.project_number == {task_number})) |= . + {
    status: $status,
    last_updated: $ts
  }' specs/state.json > specs/tmp/state.json && mv specs/tmp/state.json specs/state.json

1. Update TODO.md status marker:

- Find task entry: grep -n "^### {task_number}\." specs/TODO.md - Use Edit tool to change [OLD_STATUS] to [NEW_STATUS]

Status Mapping:

Return: JSON object with status "synced" and previous/new status fields.

Operation: postflight_update

Purpose: Set final status and link artifacts at GATE OUT checkpoint

Execution:

1. Update state.json status and timestamp:

jq --arg ts "$(date -u +%Y-%m-%dT%H:%M:%SZ)" --arg status "{target_status}" \
  '(.active_projects[] | select(.project_number == {task_number})) |= . + {
    status: $status,
    last_updated: $ts
  }' specs/state.json > specs/tmp/state.json && mv specs/tmp/state.json specs/state.json

1. Add artifacts to state.json (for each artifact):

IMPORTANT: Use two-step jq pattern to avoid Issue #1132 escaping bug. See jq-escaping-workarounds.md.

# Step 1: Update timestamp
jq --arg ts "$(date -u +%Y-%m-%dT%H:%M:%SZ)" \
  '(.active_projects[] | select(.project_number == {task_number})) |= . + {
    last_updated: $ts
  }' specs/state.json > specs/tmp/state.json && mv specs/tmp/state.json specs/state.json

# Step 2: Add artifact (append to array)
jq --arg path "{artifact_path}" \
   --arg type "{artifact_type}" \
  '(.active_projects[] | select(.project_number == {task_number})).artifacts += [{"path": $path, "type": $type}]' \
  specs/state.json > specs/tmp/state.json && mv specs/tmp/state.json specs/state.json

1. Update TODO.md status marker:

- Use Edit to change status: [RESEARCHING] -> [RESEARCHED]

1. Link artifacts in TODO.md:

- Add research/plan/summary links in appropriate location

Status Mapping:

Return: JSON object with target_status and artifacts_linked fields.

Operation: artifact_link

Purpose: Add single artifact link (idempotent)

Execution:

1. Idempotency check:

if grep -A 30 "^### {task_number}\." specs/TODO.md | grep -q "{artifact_path}"; then
  echo "Link already exists"
  # Return "skipped" status
fi

1. Add to state.json artifacts array:

IMPORTANT: Use two-step jq pattern to avoid Issue #1132 escaping bug. See jq-escaping-workarounds.md.

# Step 1: Update timestamp
jq --arg ts "$(date -u +%Y-%m-%dT%H:%M:%SZ)" \
  '(.active_projects[] | select(.project_number == {task_number})) |= . + {
    last_updated: $ts
  }' specs/state.json > specs/tmp/state.json && mv specs/tmp/state.json specs/state.json

# Step 2: Add artifact (append to array)
jq --arg path "{artifact_path}" \
   --arg type "{artifact_type}" \
  '(.active_projects[] | select(.project_number == {task_number})).artifacts += [{"path": $path, "type": $type}]' \
  specs/state.json > specs/tmp/state.json && mv specs/tmp/state.json specs/state.json

1. Add link to TODO.md using Edit tool:

Count-Aware Logic: Use inline format for 1 artifact, multi-line list for 2+. See state-management.md for detection patterns and insertion examples.

Insertion order:

• research: after Language line
• plan: after Research line (or Language if no Research)
• summary: after Plan line

Return: JSON object with status "linked" or "skipped".

Return Format

preflight_update returns:

{
  "status": "synced",
  "summary": "Updated task #{N} to [{STATUS}]",
  "previous_status": "not_started",
  "new_status": "researching"
}

postflight_update returns:

{
  "status": "{target_status}",
  "summary": "Updated task #{N} to [{STATUS}] with {M} artifacts",
  "artifacts_linked": ["path1", "path2"],
  "previous_status": "researching",
  "new_status": "researched"
}

artifact_link returns:

{
  "status": "linked|skipped",
  "summary": "Linked artifact to task #{N}" | "Link already exists",
  "artifact_path": "path/to/artifact.md",
  "artifact_type": "research"
}

Error Handling

Task Not Found

Return failed status with recommendation to verify task number.

Invalid Status Transition

Return failed status with current status and allowed transitions.

File Write Failure

Return failed status with recommendation to check permissions.

jq Parse Failure

If jq commands fail with INVALID_CHARACTER or syntax error (Issue #1132):

1. Log to errors.json with session_id and original command
2. Retry with two-step pattern from jq-escaping-workarounds.md
3. If retry succeeds, log recovery action

Integration Notes

For Workflow Commands: Do NOT use this skill directly. Workflow skills now handle their own status updates inline.

For Manual Operations: Use this skill for standalone status corrections:

### Manual Status Correction
Invoke skill-status-sync with:
- operation: preflight_update or postflight_update
- task_number: {N}
- target_status: {valid_status}
- session_id: manual_correction
- artifacts: [{path, type}, ...] (for postflight only)

This skill ensures:

• Atomic updates across both files
• Consistent jq/Edit patterns
• Proper error handling
• Direct execution without subagent overhead