Files
msd-core/get-shit-done/workflows/code-review-fix.md
Bill Huang 99c089bfbf feat: add /gsd:code-review and /gsd:code-review-fix commands (#1630)
* feat: add /gsd:code-review and /gsd:code-review-fix commands

Closes #1636

Add two new slash commands that close the gap between phase execution
and verification. After /gsd:execute-phase completes, /gsd:code-review
reviews produced code for bugs, security issues, and quality problems.
/gsd:code-review-fix then auto-fixes issues found by the review.

## New Files

- agents/gsd-code-reviewer.md — Review agent with 3 depth levels
  (quick/standard/deep) and structured REVIEW.md output
- agents/gsd-code-fixer.md — Fix agent with atomic git rollback,
  3-tier verification, per-finding atomic commits, logic-bug flagging
- commands/gsd/code-review.md — Slash command definition
- commands/gsd/code-review-fix.md — Slash command definition
- get-shit-done/workflows/code-review.md — Review orchestration:
  3-tier file scoping, repo-boundary path validation, config gate
- get-shit-done/workflows/code-review-fix.md — Fix orchestration:
  --all/--auto flags, 3-iteration cap, artifact backup across iterations
- tests/code-review.test.cjs — 35 tests covering agents, commands,
  workflows, config, integration, rollback strategy, and logic-bug flagging

## Modified Files

- get-shit-done/bin/lib/config.cjs — Register workflow.code_review and
  workflow.code_review_depth with defaults and typo suggestions
- get-shit-done/workflows/execute-phase.md — Add code_review_gate step
  (PIPE-01): runs after aggregate_results, advisory only, non-blocking
- get-shit-done/workflows/quick.md — Add Step 6.25 code review (PIPE-03):
  scopes via git diff, uses gsd-code-reviewer, advisory only
- get-shit-done/workflows/autonomous.md — Add Step 3c.5 review+fix chain
  (PIPE-02): auto-chains code-review-fix --auto when issues found

## Design Decisions

- Rollback uses git checkout -- {file} (atomic) not Write tool (partial write risk)
- Logic-bug fixes flagged "requires human verification" (syntax check cannot verify semantics)
- Path traversal guard rejects --files paths outside repo root
- Fail-closed scoping: no HEAD~N heuristics when scope is ambiguous

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* feat: add /gsd:code-review and /gsd:code-review-fix commands

Closes #1636

Add two new slash commands that close the gap between phase execution
and verification. After /gsd:execute-phase completes, /gsd:code-review
reviews produced code for bugs, security issues, and quality problems.
/gsd:code-review-fix then auto-fixes issues found by the review.

## New Files

- agents/gsd-code-reviewer.md — Review agent: 3 depth levels, REVIEW.md
- agents/gsd-code-fixer.md — Fix agent: git rollback, 3-tier verification,
  logic-bug flagging, per-finding atomic commits
- commands/gsd/code-review.md, code-review-fix.md — Slash command definitions
- get-shit-done/workflows/code-review.md — Review orchestration: 3-tier
  file scoping, path traversal guard, config gate
- get-shit-done/workflows/code-review-fix.md — Fix orchestration:
  --all/--auto flags, 3-iteration cap, artifact backup
- tests/code-review.test.cjs — 35 tests: agents, commands, workflows,
  config, integration, rollback, logic-bug flagging

## Modified Files

- get-shit-done/bin/lib/config.cjs — Register workflow.code_review and
  workflow.code_review_depth config keys
- get-shit-done/workflows/execute-phase.md — Add code_review_gate step
  (PIPE-01): after aggregate_results, advisory, non-blocking
- get-shit-done/workflows/quick.md — Add Step 6.25 code review (PIPE-03):
  git diff scoping, gsd-code-reviewer, advisory
- get-shit-done/workflows/autonomous.md — Add Step 3c.5 review+fix chain
  (PIPE-02): auto-chains code-review-fix --auto when issues found

## Design decisions

- Rollback uses git checkout -- {file} (atomic) not Write tool
- Logic-bug fixes flagged requires human verification (syntax != semantics)
- --files paths validated within repo root (path traversal guard)
- Fail-closed: no HEAD~N heuristics when scope ambiguous

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: resolve contradictory rollback instructions in gsd-code-fixer

rollback_strategy said git checkout, critical_rules said Write tool.
Align all three sections (rollback_strategy, execution_flow step b,
critical_rules) to use git checkout -- {file} consistently.

Also remove in-memory PRE_FIX_CONTENT capture — no longer needed
since git checkout is the rollback mechanism.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: address all review feedback from rounds 3-4

Blocking (bash compatibility):
- Replace mapfile -t with portable while IFS= read -r loops in both
  workflows (mapfile is bash 4+; macOS ships bash 3.2 by default)
- Add macOS bash version note to platform_notes

Blocking (quick.md scope heuristic):
- Replace fragile HEAD~$(wc -l SUMMARY.md) with git log --grep based
  diff, matching the more robust approach in code-review.md

Security (path traversal):
- Document realpath -m macOS behavior in platform_notes; guard remains
  fail-closed on macOS without coreutils

Logic / correctness:
- Fix REVIEW_PATH / FIX_REPORT_PATH interpolation in node -e strings;
  use process.env.REVIEW_PATH via env var prefix to avoid single-quote
  path injection risk
- Add iteration semantics comment clarifying off-by-one behavior
- Remove duplicate "3. Determine changed files" heading in gsd-code-reviewer.md

Agent:
- Add logic-bug limitation section to gsd-code-fixer verification_strategy

Tests (39 total, up from 32):
- Add rollback uses git checkout test
- Add success_criteria consistency test (must not say Write tool)
- Add logic-bug flagging test
- Add files_reviewed_list spec test
- Add path traversal guard structural test
- Add mapfile-in-bash-blocks tests (bash 3.2 compatibility)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: add gsd-code-reviewer to quick.md available_agent_types and copilot install test

- quick.md Step 6.25 spawns gsd-code-reviewer but the workflow's
  <available_agent_types> block did not list it, failing the spawn
  consistency CI check (#1357)
- copilot-install.test.cjs hardcoded agent list was missing
  gsd-code-fixer.agent.md and gsd-code-reviewer.agent.md, failing
  the Copilot full install verification test

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: replace /gsd: colon refs with /gsd- hyphen format in new files

Fixes stale-colon-refs CI test (#1748). All 19 violations replaced:
- agents/gsd-code-fixer.md (2): description + role spawned-by text
- agents/gsd-code-reviewer.md (4): description + role + fallback note + error msg
- get-shit-done/workflows/code-review-fix.md (7): error msgs + retry suggestions
- get-shit-done/workflows/code-review.md (5): error msgs + retry suggestions
- get-shit-done/workflows/execute-phase.md (1): code_review_gate suggestion

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-05 19:43:45 -04:00

18 KiB

Auto-fix issues from REVIEW.md. Validates phase, checks config gate, verifies REVIEW.md exists and has fixable issues, spawns gsd-code-fixer agent, handles --auto iteration loop (capped at 3), commits REVIEW-FIX.md once at the end, and presents results.

<required_reading> Read all files referenced by the invoking prompt's execution_context before starting. </required_reading>

<available_agent_types>

  • gsd-code-fixer: Applies fixes to code review findings
  • gsd-code-reviewer: Reviews source files for bugs and issues </available_agent_types>
Parse arguments and load project state:
PHASE_ARG="${1}"
INIT=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" init phase-op "${PHASE_ARG}")
if [[ "$INIT" == @file:* ]]; then INIT=$(cat "${INIT#@file:}"); fi

Parse from init JSON: phase_found, phase_dir, phase_number, phase_name, padded_phase, commit_docs.

Input sanitization (defense-in-depth):

# Validate PADDED_PHASE contains only digits and optional dot (e.g., "02", "03.1")
if ! [[ "$PADDED_PHASE" =~ ^[0-9]+(\.[0-9]+)?$ ]]; then
  echo "Error: Invalid phase number format: '${PADDED_PHASE}'. Expected digits (e.g., 02, 03.1)."
  # Exit workflow
fi

Phase validation (before config gate): If phase_found is false, report error and exit:

Error: Phase ${PHASE_ARG} not found. Run /gsd-status to see available phases.

This runs BEFORE config gate check so user errors are surfaced immediately regardless of config state.

Parse optional flags from $ARGUMENTS:

FIX_ALL=false
AUTO_MODE=false
for arg in "$@"; do
  if [[ "$arg" == "--all" ]]; then FIX_ALL=true; fi
  if [[ "$arg" == "--auto" ]]; then AUTO_MODE=true; fi
done

Compute scope variable:

if [ "$FIX_ALL" = "true" ]; then
  FIX_SCOPE="all"
else
  FIX_SCOPE="critical_warning"
fi

Compute review and fix report paths:

REVIEW_PATH="${PHASE_DIR}/${PADDED_PHASE}-REVIEW.md"
FIX_REPORT_PATH="${PHASE_DIR}/${PADDED_PHASE}-REVIEW-FIX.md"
Check if code review is enabled via config:
CODE_REVIEW_ENABLED=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" config-get workflow.code_review 2>/dev/null || echo "true")

If CODE_REVIEW_ENABLED is "false":

Code review fix skipped (workflow.code_review=false in config)

Exit workflow.

Default is true — only skip on explicit false. This check runs AFTER phase validation so invalid phase errors are shown first.

Note: This reuses the workflow.code_review config key rather than introducing a separate workflow.code_review_fix key. Rationale: fixes are meaningless without review, so a single toggle makes sense. If independent control is needed later, a separate key can be added in v2.

Verify that REVIEW.md exists:
if [ ! -f "${REVIEW_PATH}" ]; then
  echo "Error: No REVIEW.md found for Phase ${PHASE_ARG}. Run /gsd-code-review ${PHASE_ARG} first."
  exit 1
fi

Do NOT auto-run code-review. Require explicit user action to ensure review intent is clear.

Parse REVIEW.md frontmatter to check status and extract context for --auto loop:
# Parse status field
REVIEW_STATUS=$(REVIEW_PATH="${REVIEW_PATH}" node -e "
  const fs = require('fs');
  const content = fs.readFileSync(process.env.REVIEW_PATH, 'utf-8');
  const match = content.match(/^---\n([\s\S]*?)\n---/);
  if (match && /status:\s*(\S+)/.test(match[1])) {
    console.log(match[1].match(/status:\s*(\S+)/)[1]);
  } else {
    console.log('unknown');
  }
" 2>/dev/null)

If status is "clean" or "skipped":

No issues to fix in Phase ${PHASE_ARG} REVIEW.md (status: ${REVIEW_STATUS}).

Exit workflow.

If status is "unknown":

Warning: Could not parse REVIEW.md status. Proceeding with fix attempt.

Extract review depth for --auto re-review:

REVIEW_DEPTH=$(REVIEW_PATH="${REVIEW_PATH}" node -e "
  const fs = require('fs');
  const content = fs.readFileSync(process.env.REVIEW_PATH, 'utf-8');
  const match = content.match(/^---\n([\s\S]*?)\n---/);
  if (match && /depth:\s*(\S+)/.test(match[1])) {
    console.log(match[1].match(/depth:\s*(\S+)/)[1]);
  } else {
    console.log('standard');
  }
" 2>/dev/null)

Extract original review file list for --auto re-review scope persistence:

# Extract review file list — portable bash 3.2+ (no mapfile, handles spaces in paths)
REVIEW_FILES_ARRAY=()
while IFS= read -r line; do
  [ -n "$line" ] && REVIEW_FILES_ARRAY+=("$line")
done < <(REVIEW_PATH="${REVIEW_PATH}" node -e "
  const fs = require('fs');
  const content = fs.readFileSync(process.env.REVIEW_PATH, 'utf-8');
  const match = content.match(/^---\n([\s\S]*?)\n---/);
  if (match) {
    const fm = match[1];
    // Try YAML array format: files_reviewed_list: [file1, file2]
    const bracketMatch = fm.match(/files_reviewed_list:\s*\[([^\]]+)\]/);
    if (bracketMatch) {
      bracketMatch[1].split(',').map(f => f.trim()).filter(Boolean).forEach(f => console.log(f));
    } else {
      // Try YAML list format: files_reviewed_list:\n  - file1\n  - file2
      let inList = false;
      for (const line of fm.split('\n')) {
        if (/files_reviewed_list:/.test(line)) { inList = true; continue; }
        if (inList && /^\s+-\s+(.+)/.test(line)) { console.log(line.match(/^\s+-\s+(.+)/)[1].trim()); }
        else if (inList && /^\S/.test(line)) { break; }
      }
    }
  }
" 2>/dev/null)

If REVIEW.md contains a files_reviewed_list frontmatter field, use that as the re-review scope. If not present, fall back to re-reviewing the full phase (same behavior as initial code-review).

Spawn the gsd-code-fixer agent with config:
# Build config for agent
echo "Applying fixes from ${REVIEW_PATH}..."
echo "Fix scope: ${FIX_SCOPE}"

Use Task() to spawn agent:

Task(subagent_type="gsd-code-fixer", prompt="
<files_to_read>
${REVIEW_PATH}
</files_to_read>

<config>
phase_dir: ${PHASE_DIR}
padded_phase: ${PADDED_PHASE}
review_path: ${REVIEW_PATH}
fix_scope: ${FIX_SCOPE}
fix_report_path: ${FIX_REPORT_PATH}
iteration: 1
</config>

Read REVIEW.md findings, apply fixes, commit each atomically, write REVIEW-FIX.md. Do NOT commit REVIEW-FIX.md (orchestrator handles that).
")

Agent failure handling:

If Task() fails:

Error: Code fix agent failed: ${error_message}

Check if FIX_REPORT_PATH exists:

  • If yes: "Partial success — some fixes may have been committed."
  • If no: "No fixes applied."

Either way:

Some fix commits may already exist in git history — check git log for fix(${PADDED_PHASE}) commits.
You can retry with /gsd-code-review-fix ${PHASE_ARG}.

Exit workflow (skip auto loop).

Only runs if AUTO_MODE is true. If AUTO_MODE is false, skip this step entirely.
if [ "$AUTO_MODE" = "true" ]; then
  # Iteration semantics: the initial fix pass (step 5) is iteration 1.
  # This loop runs iterations 2..MAX_ITERATIONS (re-review + re-fix cycles).
  # Total fix passes = MAX_ITERATIONS. Loop uses -lt (not -le) intentionally.
  ITERATION=1
  MAX_ITERATIONS=3
  
  while [ $ITERATION -lt $MAX_ITERATIONS ]; do
    ITERATION=$((ITERATION + 1))
    
    echo ""
    echo "═══════════════════════════════════════════════════════"
    echo "  --auto: Starting iteration ${ITERATION}/${MAX_ITERATIONS}"
    echo "═══════════════════════════════════════════════════════"
    echo ""
    
    # Re-review using same depth and file scope as original review
    echo "Re-reviewing phase ${PHASE_ARG} at ${REVIEW_DEPTH} depth..."
    
    # Backup previous REVIEW.md and REVIEW-FIX.md before overwriting
    if [ -f "${REVIEW_PATH}" ]; then
      cp "${REVIEW_PATH}" "${REVIEW_PATH%.md}.iter${ITERATION}.md" 2>/dev/null || true
    fi
    if [ -f "${FIX_REPORT_PATH}" ]; then
      cp "${FIX_REPORT_PATH}" "${FIX_REPORT_PATH%.md}.iter${ITERATION}.md" 2>/dev/null || true
    fi
    
    # If original review had explicit file list, pass it safely to re-review agent
    FILES_CONFIG=""
    if [ ${#REVIEW_FILES_ARRAY[@]} -gt 0 ]; then
      FILES_CONFIG="files:"
      for f in "${REVIEW_FILES_ARRAY[@]}"; do
        FILES_CONFIG="${FILES_CONFIG}
  - ${f}"
      done
    fi
    
    # Spawn gsd-code-reviewer agent to re-review
    # (This overwrites REVIEW_PATH with latest review state)
    Task(subagent_type="gsd-code-reviewer", prompt="
<config>
depth: ${REVIEW_DEPTH}
phase_dir: ${PHASE_DIR}
review_path: ${REVIEW_PATH}
${FILES_CONFIG}
</config>

Re-review the phase at ${REVIEW_DEPTH} depth. Write findings to ${REVIEW_PATH}.
Do NOT commit the output — the orchestrator handles that.
")
    
    # Check new REVIEW.md status
    NEW_STATUS=$(REVIEW_PATH="${REVIEW_PATH}" node -e "
      const fs = require('fs');
      const content = fs.readFileSync(process.env.REVIEW_PATH, 'utf-8');
      const match = content.match(/^---\n([\s\S]*?)\n---/);
      if (match && /status:\s*(\S+)/.test(match[1])) {
        console.log(match[1].match(/status:\s*(\S+)/)[1]);
      } else {
        console.log('unknown');
      }
    " 2>/dev/null)
    
    if [ "$NEW_STATUS" = "clean" ]; then
      echo ""
      echo "✓ All issues resolved after iteration ${ITERATION}."
      break
    fi
    
    # Still has issues — spawn fixer again
    echo "Issues remain. Applying fixes for iteration ${ITERATION}..."
    
    Task(subagent_type="gsd-code-fixer", prompt="
<files_to_read>
${REVIEW_PATH}
</files_to_read>

<config>
phase_dir: ${PHASE_DIR}
padded_phase: ${PADDED_PHASE}
review_path: ${REVIEW_PATH}
fix_scope: ${FIX_SCOPE}
fix_report_path: ${FIX_REPORT_PATH}
iteration: ${ITERATION}
</config>

Read REVIEW.md findings, apply fixes, commit each atomically, write REVIEW-FIX.md (overwrite previous). Do NOT commit REVIEW-FIX.md.
")
    
    # Check if fixer succeeded
    if [ ! -f "${FIX_REPORT_PATH}" ]; then
      echo "Warning: Iteration ${ITERATION} fixer failed to produce fix report. Stopping auto-loop."
      break
    fi
  done
  
  # After loop completes
  if [ $ITERATION -ge $MAX_ITERATIONS ]; then
    echo ""
    echo "⚠ Reached maximum iterations (${MAX_ITERATIONS}). Remaining issues documented in REVIEW-FIX.md."
  fi
fi

Key design decisions for --auto (addresses ALL review HIGH concerns):

  1. Re-review scope: Uses REVIEW_FILES_ARRAY from original REVIEW.md frontmatter, falling back to full phase scope. Scope is NOT lost between iterations. Uses portable while-read loop (bash 3.2+ compatible, handles spaces in paths).
  2. Artifact semantics: REVIEW.md is overwritten by each re-review (latest review state). REVIEW-FIX.md is overwritten by each fixer iteration (latest fix state with iteration count). There is ONE final version of each artifact, not per-iteration copies. Backup files (.iterN.md) preserve history for post-mortem analysis if iterations degrade.
  3. Commit timing: Fix commits happen per-finding inside the agent. REVIEW-FIX.md is NOT committed until step 7 (after ALL iterations complete). Only ONE docs commit for REVIEW-FIX.md, not one per iteration.
After ALL iterations complete (or single pass in non-auto mode), validate and commit REVIEW-FIX.md:
if [ -f "${FIX_REPORT_PATH}" ]; then
  # Validate REVIEW-FIX.md has valid YAML frontmatter with status field
  HAS_STATUS=$(REVIEW_PATH="${REVIEW_PATH}" node -e "
    const fs = require('fs');
    const content = fs.readFileSync(process.env.FIX_REPORT_PATH, 'utf-8');
    const match = content.match(/^---\n([\s\S]*?)\n---/);
    if (match && /status:/.test(match[1])) { console.log('valid'); } else { console.log('invalid'); }
  " 2>/dev/null)
  
  if [ "$HAS_STATUS" = "valid" ]; then
    echo "REVIEW-FIX.md created at ${FIX_REPORT_PATH}"
    
    if [ "$COMMIT_DOCS" = "true" ]; then
      node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" commit \
        "docs(${PADDED_PHASE}): add code review fix report" \
        --files "${FIX_REPORT_PATH}"
    fi
  else
    echo "Warning: REVIEW-FIX.md has invalid frontmatter (no status field). Not committing."
    echo "Agent may have produced malformed output. Review manually: ${FIX_REPORT_PATH}"
  fi
else
  echo "Warning: REVIEW-FIX.md not found at ${FIX_REPORT_PATH}."
  echo "Agent may have failed before writing report."
  echo "Check git log for any fix(${PADDED_PHASE}) commits that were applied."
fi

This commit happens ONCE at the end of the workflow, after all iterations (if --auto) complete. Not per-iteration.

Parse REVIEW-FIX.md frontmatter and present formatted summary to user.

First check if fix report exists:

if [ ! -f "${FIX_REPORT_PATH}" ]; then
  echo ""
  echo "═══════════════════════════════════════════════════════════════"
  echo ""
  echo "  ⚠ No fix report generated"
  echo ""
  echo "───────────────────────────────────────────────────────────────"
  echo ""
  echo "The fixer agent may have failed before completing."
  echo "Check git log for any fix(${PADDED_PHASE}) commits."
  echo ""
  echo "Retry: /gsd-code-review-fix ${PHASE_ARG}"
  echo ""
  echo "═══════════════════════════════════════════════════════════════"
  exit 1
fi

Extract frontmatter fields:

# Extract only the YAML frontmatter block (between first two --- lines)
FIX_FRONTMATTER=$(REVIEW_PATH="${REVIEW_PATH}" node -e "
  const fs = require('fs');
  const content = fs.readFileSync(process.env.FIX_REPORT_PATH, 'utf-8');
  const match = content.match(/^---\n([\s\S]*?)\n---/);
  if (match) process.stdout.write(match[1]);
" 2>/dev/null)

# Parse fields from frontmatter only (not full file)
FIX_STATUS=$(echo "$FIX_FRONTMATTER" | grep "^status:" | cut -d: -f2 | xargs)
FINDINGS_IN_SCOPE=$(echo "$FIX_FRONTMATTER" | grep "^findings_in_scope:" | cut -d: -f2 | xargs)
FIXED_COUNT=$(echo "$FIX_FRONTMATTER" | grep "^fixed:" | cut -d: -f2 | xargs)
SKIPPED_COUNT=$(echo "$FIX_FRONTMATTER" | grep "^skipped:" | cut -d: -f2 | xargs)
ITERATION_COUNT=$(echo "$FIX_FRONTMATTER" | grep "^iteration:" | cut -d: -f2 | xargs)

Display formatted inline summary:

echo ""
echo "═══════════════════════════════════════════════════════════════"
echo ""
echo "  Code Review Fix Complete: Phase ${PHASE_NUMBER} (${PHASE_NAME})"
echo ""
echo "───────────────────────────────────────────────────────────────"
echo ""
echo "  Fix Scope:       ${FIX_SCOPE}"
echo "  Findings:        ${FINDINGS_IN_SCOPE}"
echo "  Fixed:           ${FIXED_COUNT}"
echo "  Skipped:         ${SKIPPED_COUNT}"
if [ "$AUTO_MODE" = "true" ]; then
  echo "  Iterations:      ${ITERATION_COUNT}"
fi
echo "  Status:          ${FIX_STATUS}"
echo ""
echo "───────────────────────────────────────────────────────────────"
echo ""

If status is "all_fixed":

if [ "$FIX_STATUS" = "all_fixed" ]; then
  echo "✓ All issues resolved."
  echo ""
  echo "Full report: ${FIX_REPORT_PATH}"
  echo ""
  echo "Next step:"
  echo "  /gsd-verify-work  — Verify phase completion"
  echo ""
fi

If status is "partial" or "none_fixed":

if [ "$FIX_STATUS" = "partial" ] || [ "$FIX_STATUS" = "none_fixed" ]; then
  echo "⚠ Some issues could not be fixed automatically."
  echo ""
  echo "Full report: ${FIX_REPORT_PATH}"
  echo ""
  echo "Next steps:"
  echo "  cat ${FIX_REPORT_PATH}                     — View fix report"
  echo "  /gsd-code-review ${PHASE_NUMBER}           — Re-review code"
  echo "  /gsd-verify-work                           — Verify phase completion"
  echo ""
fi
echo "═══════════════════════════════════════════════════════════════"

<platform_notes> Windows: This workflow uses bash features (arrays, variable expansion, while loops). On Windows, it requires Git Bash or WSL. Native PowerShell is not supported. The CI matrix (Ubuntu/macOS/Windows) runs under Git Bash on Windows runners, which provides bash compatibility. </platform_notes>

<success_criteria>

  • Phase validated before config gate check
  • Config gate checked (workflow.code_review)
  • REVIEW.md existence verified (error if missing)
  • REVIEW.md status checked (skip if clean/skipped)
  • Agent spawned with correct config (review_path, fix_scope, fix_report_path)
  • Agent failure handled with partial-success awareness (some fix commits may exist)
  • --auto iteration loop respects 3-iteration cap
  • --auto re-review uses persisted file scope (not lost between iterations)
  • REVIEW-FIX.md committed ONCE after all iterations (not per-iteration)
  • Missing fix report handled with explicit error message in present_results
  • Results presented inline with next step suggestion </success_criteria>