deep-receipt

Invocation

이 스킬은 두 가지 경로로 호출됩니다 — 어느 쪽이든 본 SKILL 본문의 절차를 그대로 실행합니다:

1. Claude Code 슬래시 — 사용자가 /deep-receipt [args...] 입력 (skill 의 user-invocable: true 가 슬래시 진입을 허용).
2. 타 에이전트 / Codex / Copilot CLI / Gemini CLI / SDK — Skill({ skill: "deep-work:deep-receipt", args: "..." }) 형태로 명시 invoke (cross-platform 표준 경로).

두 경로 모두 args 는 동일한 토큰 문자열로 전달되며, 본문 ($ARGUMENTS 자리) 의 파서가 동일하게 처리합니다.

Inputs (skill args)

인자	의미
(없음) / dashboard	ASCII visual dashboard of all slice receipts
view SLICE-NNN	특정 slice 의 receipt 상세
export	JSON / Markdown export
validate	8-item 검증 (verify-delegated-receipt-runner)

빈 args / 매칭되지 않는 토큰 → 본문의 default 분기로 진입.

Prerequisites

이 entry skill 은 deep-work-orchestrator (Phase dispatch) 및 deep-work-workflow (reference skill — Phase 규약/Exit Gate/M3 envelope) 와 함께 동작합니다. 활성 deep-work 세션이 있을 때는 세션 state file (.claude/deep-work.<SESSION_ID>.md) 의 변수 (work_dir, current_phase, active_slice 등) 를 읽어 동작하며, 세션 외부에서도 standalone 실행이 가능한 경우 본문의 분기를 따릅니다.

Hub-spoke 관계: 본 skill 은 deep-status hub 의 sub-page 입니다 — /deep-status --<flag> 가 본문 로직을 inline Read 하여 실행하는 것이 주 경로이며, 직접 호출도 동일하게 지원됩니다.

Cross-platform self-containment: Claude Code 에서는 sibling skill 이 description 매칭으로 자동 로드됩니다. Codex / Copilot CLI / Gemini CLI / Agent SDK 에서 Skill() 로 호출 시 sibling auto-load 보장이 약할 수 있으므로, 본문은 self-contained 으로 보존되어 있습니다 — state file 해석, $ARGUMENTS 파싱, AskUserQuestion 분기, 출력 포맷이 인라인.

Internal (v6.3.0) — /deep-status --receipts가 이 파일의 display logic을 Read하여 실행합니다. 자동 호출이 주 경로이며, 직접 호출도 지원됩니다. 참조처: skills/deep-status/SKILL.md §6 (Read skills/deep-receipt/SKILL.md and follow its display logic inline).

Receipt Management (v4.0)

View, export, and manage evidence receipts from the implementation phase.

Language

Detect the user's language from their messages or the Claude Code language setting. Output ALL user-facing messages in the detected language.

Usage

• /deep-receipt — Show receipt dashboard (same as /deep-receipt dashboard)
• /deep-receipt dashboard — ASCII visual dashboard of all slice receipts
• /deep-receipt view SLICE-NNN — Show detailed receipt for a specific slice
• /deep-receipt export --format=json — Export all receipts as single JSON file
• /deep-receipt export --format=md — Export as markdown (for PR descriptions)
• /deep-receipt export --format=ci — Export CI bundle (session-receipt + all slice receipts in one JSON, for GitHub Actions validation)

Runtime Setup

Resolve the current session's state file:

1. If DEEP_WORK_SESSION_ID env var is set → .claude/deep-work.${DEEP_WORK_SESSION_ID}.md
2. If .claude/deep-work-current-session pointer file exists → read session ID → .claude/deep-work.${SESSION_ID}.md
3. Legacy fallback → .claude/deep-work.local.md

Set $STATE_FILE to the resolved path.

Read $STATE_FILE and extract work_dir. Receipts are stored in $WORK_DIR/receipts/SLICE-NNN.json.

Envelope-aware reads (v6.5.0): deep-work 6.5.0 부터 SLICE-*.json 과 session-receipt.json 는 M3 cross-plugin envelope ({schema_version: "1.0", envelope: {...}, payload: {...}}) 로 emit 된다 (cf. claude-deep-suite/docs/envelope-migration.md §1). 본 명령의 모든 receipt read 단계에서 다음 unwrap 규칙을 적용한다:

1. JSON 파싱 후 root 가 envelope 형태이면 (schema_version === "1.0", envelope 객체, payload 키 모두 존재) identity guard 검증: envelope.producer === "deep-work" ∧ envelope.artifact_kind ∈ {slice-receipt, session-receipt} ∧ envelope.schema.name === envelope.artifact_kind. 위 조건을 모두 통과하면 payload 객체를 사용 (legacy receipt body). 한 항목이라도 어긋나면 "foreign envelope at " 경고 후 receipt 무시 (handoff §4 round-4 identity guard).
2. Legacy(non-envelope) JSON 이면 그대로 사용 (forward-compat).
3. /deep-receipt export --format=ci 같은 bundle export 는 envelope wrapping 을 그대로 유지한 채 묶어 외부 CI 에 전달한다 (envelope 의 run_id chain 이 audit 용).

Dashboard

Scan $WORK_DIR/receipts/ directory for all receipt JSON files. For each receipt, display:

Receipt Dashboard

┌──────────┬──────────┬──────────┬──────────┬──────────┬──────────┐
│ Slice    │ TDD      │ Tests    │ Spec     │ Contract │ Review   │
├──────────┼──────────┼──────────┼──────────┼──────────┼──────────┤
│ SLICE-001│ ✅ GREEN │ 5/5 PASS │ 3/3 ✅   │ 4/4 PASS │ PASS     │
│ SLICE-002│ 🟡 RED_V │ 2/5 FAIL │ 1/3 ⏳   │ 1/3 FAIL │ —        │
│ SLICE-003│ ⬜ PEND  │ —        │ —        │ —        │ —        │
│ SLICE-004│ SPIKE    │ —        │ —        │ —        │ ⚠️       │
└──────────┴──────────┴──────────┴──────────┴──────────┴──────────┘

요약:
  완료: 1/4 (25%)
  TDD 준수: 1 strict, 0 relaxed, 0 override, 1 spike
  총 변경: +142 -23 (8 files)

TDD state icons:

• ✅ GREEN/REFACTOR — TDD cycle complete
• 🔴 RED — failing test written, not yet green
• 🟡 RED_VERIFIED — verified failing test, implementation pending
• 🟢 GREEN_ELIGIBLE — production code written, verification pending
• ⬜ PENDING — not started
• SPIKE — spike mode (not merge-eligible)
• override — TDD skipped by user (merge-eligible with warning)
• 🔍 SENSOR_RUN — Computational sensor running
• 🔧 SENSOR_FIX — Fixing sensor errors (self-correction loop active)
• ✅ SENSOR_CLEAN — All sensors passed

Health Check Display

When displaying any receipt (dashboard or view), if the session state file contains health_report, include the Health Check section:

### Health Check (Phase 1 진단)
- 🔍 드리프트: dead-export {count}건 | coverage {delta}%p | vuln {critical+high}건 | stale {count}건
- 📐 Fitness: {passed}/{total} 통과 | 위반 delta: {delta}건
- ⚠️ Required: {acknowledged ? "acknowledged" : "미해결 N건"}

Steps:

1. Read health_report from the session state file
2. Extract drift metrics: health_report.drift.dead_exports.count, health_report.drift.coverage_trend.delta, health_report.drift.dependency_vuln.critical + health_report.drift.dependency_vuln.high, health_report.drift.stale_config.count
3. Extract fitness metrics: health_report.fitness.passed, health_report.fitness.total_rules, health_report.fitness.required_missing
4. Extract required status: check acknowledged_required_issues in the state file
5. If health_report is absent from the state file, skip this section silently

View

/deep-receipt view SLICE-NNN:

Read $WORK_DIR/receipts/SLICE-NNN.json and display formatted:

Receipt: SLICE-NNN — [Goal from plan.md]

TDD Cycle:
  🔴 RED:   [timestamp] — [test name]
  🟢 GREEN: [timestamp] — [N tests passing]

Changes:
  Files: [file1, file2]
  Diff:  +[N] -[N] lines

Spec Compliance:
  ✅ [requirement 1]
  ✅ [requirement 2]
  ❌ [requirement 3]

Contract Compliance:
  ✅ [contract item 1]
  ✅ [contract item 2]
  ❌ [contract item 3]
  Threshold: [all/majority]
  Result: [PASS/FAIL]

Code Review:
  결과: [PASS/WARN]
  Findings: [N] (critical: [N], important: [N])

Sensor Results:
  생태계: [ecosystem, e.g. typescript]
  Lint ([tool]): [pass|fail|not_applicable] — errors: [N], warnings: [N], correction_rounds: [N]
  Typecheck ([tool]): [pass|fail|not_applicable] — errors: [N], correction_rounds: [N]
  Coverage ([tool]): [pass|fail|not_applicable] — line: [N]%, branch: [N]%

Debug Log:
  [None / RC-NNN: root cause description]

If sensor_results is absent from the receipt, skip the Sensor Results block silently.

Export — JSON

/deep-receipt export --format=json:

Read all receipt files, combine into a single JSON array, and write to $WORK_DIR/receipts-export.json:

{
  "session": {
    "task": "[task_description]",
    "branch": "[git_branch]",
    "timestamp": "[ISO]",
    "tdd_mode": "[strict/relaxed/spike]"
  },
  "summary": {
    "total_slices": N,
    "completed": N,
    "tdd_compliance": { "strict": N, "relaxed": N, "override": N, "spike": N },
    "total_changes": { "added": N, "removed": N, "files": N },
    "debug_count": N,
    "contract_compliance": {
      "total_items": N,
      "passed": N,
      "failed": N,
      "pass_rate": "N%"
    }
  },
  "slices": [
    {
      "...other receipt fields...",
      "contract_compliance": {
        "items": { "item1": true, "item2": true, "item3": false },
        "threshold": "all",
        "result": "FAIL"
      }
    }
  ]
}

Display: Exported: $WORK_DIR/receipts-export.json

Export — Markdown

/deep-receipt export --format=md:

Generate a markdown summary suitable for PR descriptions. Write to $WORK_DIR/receipts-export.md:

## Evidence Summary

| Slice | Goal | TDD | Tests | Spec | Contract | Review |
|-------|------|-----|-------|------|----------|--------|
| SLICE-001 | [goal] | ✅ strict | 5/5 | 3/3 | 4/4 PASS | PASS |
| SLICE-002 | [goal] | ✅ strict | 3/3 | 2/2 | 3/3 PASS | PASS |

### TDD Compliance
- Strict mode: N/N slices (100%)
- Average RED→GREEN time: [Nm]

### Changes
- Total: +[N] -[N] lines across [N] files
- Debug sessions: [N] (root causes documented)

### Spec Compliance
- All requirements met: [N/N] slices

Display: Exported: $WORK_DIR/receipts-export.md

Copy to clipboard suggestion:

PR 디스크립션에 붙여넣으려면:
   cat $WORK_DIR/receipts-export.md | pbcopy

Receipt Schema: Sensor Fields

These fields are written by the sensor infrastructure during Phase 3 (implement) and are consumed by deep-test Quality Gates (Section 4-6/4-7) and deep-review integration.

Per-slice sensor_results

{
  "ecosystem": "typescript",
  "lint": {
    "tool": "eslint",
    "status": "pass|fail|not_applicable|timeout",
    "errors": 0,
    "warnings": 0,
    "correction_rounds": 0
  },
  "typecheck": {
    "tool": "tsc",
    "status": "pass|fail|not_applicable|timeout",
    "errors": 0,
    "correction_rounds": 0
  },
  "coverage": {
    "tool": "jest",
    "status": "pass|fail|not_applicable|timeout",
    "line_pct": 87.3,
    "branch_pct": 72.1
  }
}

Session mutation_testing

Written to the session state file after Phase 4 Mutation Score gate (Section 4-7).

{
  "tool": "stryker",
  "status": "completed|not_applicable",
  "total_mutants": 45,
  "killed": 39,
  "survived": 4,
  "equivalent": 2,
  "score": 90.7,
  "auto_fix_rounds": 2,
  "remaining_survived": [
    {
      "file": "src/auth/jwt.ts",
      "line": 42,
      "mutator": "ConditionalExpression",
      "tag": "possibly_equivalent"
    }
  ]
}