recce-review

/recce-review — Data Review Orchestration

This skill orchestrates MCP health checks, auto-start recovery, tracked model handoff, sub-agent dispatch, post-review cleanup, and risk-based next-step suggestions.

Follow these steps in order.

---

Step 1: MCP Health Check

Run:

bash ${CLAUDE_PLUGIN_ROOT}/scripts/check-mcp.sh

Parse the KEY=VALUE output:

• If RUNNING=true — MCP server is healthy. Skip to Step 4.
• If RUNNING=false — MCP server is not running. Proceed to Step 2.

---

Step 2: Auto-Start MCP Server

Run:

bash ${CLAUDE_PLUGIN_ROOT}/scripts/start-mcp.sh

Parse the KEY=VALUE output:

• If STATUS=STARTED or STATUS=ALREADY_RUNNING — proceed to Step 3.
• If ERROR= appears in output — abort. Show the ERROR, MESSAGE, and FIX lines verbatim. Tell the user: "MCP server failed to start. Fix the issue above, then re-run /recce-review."

---

Step 3: Re-Check MCP After Auto-Start

Note: Skip this step if Step 1 already returned RUNNING=true.

Run:

bash ${CLAUDE_PLUGIN_ROOT}/scripts/check-mcp.sh

Parse the KEY=VALUE output:

• If RUNNING=true — proceed to Step 4.
• If RUNNING=false — abort. Tell the user: "MCP server started but health check failed. Check logs and try bash ${CLAUDE_PLUGIN_ROOT}/scripts/check-mcp.sh manually."

---

Step 4: Determine Model Scope

Run:

PROJECT_HASH=$(printf '%s' "$PWD" | md5 2>/dev/null | cut -c1-8 || printf '%s' "$PWD" | md5sum | cut -c1-8)
CHANGES_FILE="/tmp/recce-changed-${PROJECT_HASH}.txt"
if [ -f "$CHANGES_FILE" ] && [ -s "$CHANGES_FILE" ]; then
    echo "TRACKED=true"
    echo "MODEL_COUNT=$(wc -l < "$CHANGES_FILE" | tr -d ' ')"
    echo "MODELS=$(while IFS= read -r f; do basename "$f" .sql; done < "$CHANGES_FILE" | paste -sd ', ' -)"
else
    echo "TRACKED=false"
fi

Parse the output:

• If TRACKED=true — record the MODELS value (comma-separated model names). Use these in Step 5.
• If TRACKED=false — no tracked changes file exists. Do not abort. Do not ask the user for model names. The agent will use state:modified+ as a fallback selector.

---

Step 5: Dispatch Review Agent

Use the agent: tool to dispatch recce-reviewer.

If tracked models were found (Step 4 returned TRACKED=true): Include in the dispatch context:

"Changed models (from tracked file): {MODELS}. Focus review on these models using selector: {model1}+ {model2}+ (one per model from the list)."

If no tracked models (Step 4 returned TRACKED=false): Include in the dispatch context:

"No tracked changes file found. Use state:modified+ as the default selector to review all modified models."

Context passthrough: If the user's request includes any of the following, include it in the dispatch message so the reviewer can validate findings against intent:

• Stakeholder request (who asked for the change and what they asked for)
• PR description (what the change claims to do)
• Change rationale (why the change was made)

Format: Context: [stakeholder] requested '[request]'. PR says: '[description]'.

This enables the reviewer's context validation step (Step 4 in the agent workflow).

Wait for the agent to complete and capture its full output.

---

Step 6: Post-Review Cleanup (on success only)

Check if the agent's output contains ## Data Review Summary.

If YES (successful review):

Run:

PROJECT_HASH=$(printf '%s' "$PWD" | md5 2>/dev/null | cut -c1-8 || printf '%s' "$PWD" | md5sum | cut -c1-8)
rm -f "/tmp/recce-changed-${PROJECT_HASH}.txt"

This clears tracked changes so the pre-commit guard no longer warns about already-reviewed models.

If NO (agent error or incomplete review):

Do not delete the file. Tell the user: "Review did not complete successfully. Tracked changes preserved for retry. Run /recce-review again." Then STOP — do not proceed to Step 7.

---

Step 7: Next Steps Based on Risk Level

Skip this step if the review did not complete successfully (no ## Data Review Summary found in Step 6).

Parse the risk level from the agent's summary output (look for Risk level: HIGH, Risk level: MEDIUM, or Risk level: LOW).

• HIGH: "Schema breaking changes detected. Consider running /recce-check for detailed profile and query analysis before committing."
• MEDIUM: "Row count changes detected. Review the deltas above, then commit when satisfied."
• LOW: "No significant data impact detected. Looks safe to commit."