---

name: pmc-simulate-tester description: Simulate the tester workflow step-by-step with real ticket system. Executes test loading, running, verification, and merging with explicit ENTER/EXIT stage declarations. Use this skill to debug, understand, or manually run the tester workflow.

PMC Simulate Tester

Manually simulate the tester workflow step-by-step, executing shell scripts and test operations with explicit stage declarations.

Required Context

Before starting, you need:

• ticket_id: Ticket ID (e.g., T00001)
• mode: Test mode - all | single | ticket-completion (default: all)
• test_id: Specific test ID (required for mode=single)
• working_dir: Project working directory

Workflow Overview

tester.workflow
├── load_tests (shell) ─────────────→ run_tests
├── run_tests (claude) ─────────────→ verification_setup
├── verification_setup (shell) ─────→ verify
├── verify (claude) ────────────────→ merge
├── merge (shell) ──────────────────→ report
├── report (shell) ─────────────────→ terminal_success
└── terminal_error (on any failure)

Execution Instructions

For each state, declare entry and exit explicitly:

ENTER STAGE: tester.workflow.{state_name}
  [perform actions]
EXIT STAGE: tester.workflow.{state_name} → {next_state} (reason: {why})

---

State: load_tests

Type: shell Script: load-tests.sh

ENTER STAGE: tester.workflow.load_tests

Execute this logic:

1. Set paths:

   TESTS_DIR = {working_dir}/docs/3-tests/tickets/{ticket_id}
   DEFINITION_FILE = TESTS_DIR/tests-definition.json
   PROGRESS_FILE = TESTS_DIR/tests-in-progress.json
   VERIFICATION_FILE = TESTS_DIR/tests-in-verification.json
   

2. Check tests-definition.json exists:

   • If NOT exists → terminal_error

3. Remove working files if they exist:

   • Delete tests-in-progress.json
   • Delete tests-in-verification.json

4. Load and filter tests based on mode:

   Mode	Filter Logic
   all	Include all tests
   single	Include only test where id == test_id
   ticket-completion	Include only tests where ticket-completion == true

5. Write filtered tests to tests-in-progress.json:

   {
     "ticket": "{ticket_id}",
     "title": "...",
     "tests": [filtered tests]
   }
   

6. Report: Loaded {N} tests (mode: {mode})

EXIT STAGE: tester.workflow.load_tests → run_tests (loaded: {N} tests)

Error: If tests-definition.json not found:

EXIT STAGE: tester.workflow.load_tests → terminal_error (tests-definition.json not found)

---

State: run_tests

Type: claude Session: start

ENTER STAGE: tester.workflow.run_tests

Use the run-tests skill to execute tests for ticket {ticket_id}.

The run-tests skill will:

1. Read tests-in-progress.json
2. Execute each test step
3. Record trajectory (actions taken)
4. Mark passes: true/false for each test
5. Update tests-in-progress.json with results

Output JSON:

{
  "tests_executed": N,
  "tests_passed": N,
  "tests_failed": N,
  "tests_blocked": N
}

EXIT STAGE: tester.workflow.run_tests → verification_setup (executed: {N}, passed: {N})

---

State: verification_setup

Type: shell Script: verification-setup.sh

ENTER STAGE: tester.workflow.verification_setup

Execute this logic:

1. Check tests-in-progress.json exists:

   • If NOT exists → terminal_error

2. Copy to tests-in-verification.json with field resets:

   Field	Action
   passes	Reset to false
   blocked	Reset to false
   step.verified	Reset to false for each step
   step.trajectory	Keep intact (don't reset)

3. Write tests-in-verification.json

4. Report: Created tests-in-verification.json, reset verified/passes for {N} tests

EXIT STAGE: tester.workflow.verification_setup → verify (tests prepared: {N})

Error: If tests-in-progress.json not found:

EXIT STAGE: tester.workflow.verification_setup → terminal_error (tests-in-progress.json not found)

---

State: verify

Type: claude Session: start

ENTER STAGE: tester.workflow.verify

Use the verify-tests skill to verify tests for ticket {ticket_id}.

The verify-tests skill will:

1. Read tests-in-verification.json
2. Follow recorded trajectories from run phase
3. Independently verify each step
4. Mark verified: true/false for each step
5. Mark passes: true/false for each test
6. Note any discrepancies between run and verify
7. Update tests-in-verification.json with results

Output JSON:

{
  "tests_verified": N,
  "tests_passed": N,
  "tests_failed": N,
  "tests_blocked": N,
  "discrepancies": ["test_id: description of discrepancy", ...]
}

EXIT STAGE: tester.workflow.verify → merge (verified: {N}, discrepancies: {N})

---

State: merge

Type: shell Script: merge-results.sh

ENTER STAGE: tester.workflow.merge

Execute this logic:

1. Check tests-in-verification.json exists:

   • If NOT exists → terminal_error

2. If tests-results.json doesn't exist:

   • If tests-definition.json exists → copy to create results
   • Else → terminal_error

3. Merge verification results into results:

   # Pseudocode
   results_by_id = {t['id']: t for t in results_tests}
   
   for verified_test in verification_tests:
       results_by_id[verified_test['id']] = verified_test
   
   results['tests'] = list(results_by_id.values())
   

4. Write merged tests-results.json

5. Report: Merged {N} tests into tests-results.json

EXIT STAGE: tester.workflow.merge → report (merged: {N} tests)

Error: If required files not found:

EXIT STAGE: tester.workflow.merge → terminal_error (missing files)

---

State: report

Type: shell Script: report.sh

ENTER STAGE: tester.workflow.report

Execute this logic:

1. Read tests-in-verification.json

2. Calculate statistics:

   total = len(tests)
   passed = count where passes == true
   blocked = count where blocked == true
   failed = total - passed
   
   # Ticket-completion subset
   tc_tests = tests where ticket-completion == true
   tc_total = len(tc_tests)
   tc_passed = count where passes == true
   tc_blocked = count where blocked == true
   

3. Identify test lists:

   failed_tests = [id for tests where passes==false and blocked==false]
   blocked_tests = [id for tests where blocked==true]
   

4. Generate report:

   {
     "total": N,
     "passed": N,
     "failed": N,
     "blocked": N,
     "ticket_completion_total": N,
     "ticket_completion_passed": N,
     "ticket_completion_blocked": N,
     "failed_tests": ["id", ...],
     "blocked_tests": ["id", ...],
     "all_passed": true|false,
     "ticket_completion_done": true|false
   }
   

5. Print report JSON

EXIT STAGE: tester.workflow.report → terminal_success

---

Terminal States

ENTER STAGE: tester.workflow.terminal_success
  Status: success
  [Workflow completed successfully]
EXIT STAGE: tester.workflow.terminal_success (workflow complete)

ENTER STAGE: tester.workflow.terminal_error
  Status: failure
  Message: "Tester workflow failed"
EXIT STAGE: tester.workflow.terminal_error (workflow failed)

---

File Flow Diagram

tests-definition.json (READ-ONLY source)
        │
        ▼ load_tests (filter by mode)
tests-in-progress.json
        │
        ▼ run_tests (execute, record trajectory)
tests-in-progress.json (updated with trajectory, passes)
        │
        ▼ verification_setup (copy, reset passes/verified)
tests-in-verification.json
        │
        ▼ verify (follow trajectory, verify independently)
tests-in-verification.json (updated with verified, passes)
        │
        ▼ merge (merge into results by test id)
tests-results.json (final merged results)
        │
        ▼ report (generate summary)
[JSON report output]

---

Example Simulation Output

=== TESTER WORKFLOW SIMULATION ===
Context: ticket_id=T00001, mode=single, test_id=T00001-03, working_dir=/project

ENTER STAGE: tester.workflow.load_tests
  TESTS_DIR: /project/docs/3-tests/tickets/T00001
  tests-definition.json exists: YES
  Removing tests-in-progress.json: DONE
  Removing tests-in-verification.json: DONE
  Mode: single, test_id: T00001-03
  Filtered: 1 test
  Writing tests-in-progress.json
EXIT STAGE: tester.workflow.load_tests → run_tests (loaded: 1 tests)

ENTER STAGE: tester.workflow.run_tests
  Using run-tests skill...
  Executing test T00001-03: "User can save document"
  Step 1: Open application → DONE (trajectory recorded)
  Step 2: Create new document → DONE (trajectory recorded)
  Step 3: Click save button → DONE (trajectory recorded)
  Step 4: Verify file exists → DONE (trajectory recorded)
  Test result: PASSED
EXIT STAGE: tester.workflow.run_tests → verification_setup (executed: 1, passed: 1)

ENTER STAGE: tester.workflow.verification_setup
  tests-in-progress.json exists: YES
  Creating tests-in-verification.json
  Reset passes=false, verified=false for 1 test
  Keeping trajectory intact
EXIT STAGE: tester.workflow.verification_setup → verify (tests prepared: 1)

ENTER STAGE: tester.workflow.verify
  Using verify-tests skill...
  Following trajectory for T00001-03...
  Step 1: Open application → verified: YES
  Step 2: Create new document → verified: YES
  Step 3: Click save button → verified: YES
  Step 4: Verify file exists → verified: YES
  Test result: PASSED (verified independently)
EXIT STAGE: tester.workflow.verify → merge (verified: 1, discrepancies: 0)

ENTER STAGE: tester.workflow.merge
  tests-in-verification.json exists: YES
  tests-results.json exists: YES
  Merging 1 test by id
  Writing tests-results.json
EXIT STAGE: tester.workflow.merge → report (merged: 1 tests)

ENTER STAGE: tester.workflow.report
  Reading tests-in-verification.json
  Report:
  {
    "total": 1,
    "passed": 1,
    "failed": 0,
    "blocked": 0,
    "ticket_completion_total": 1,
    "ticket_completion_passed": 1,
    "all_passed": true,
    "ticket_completion_done": true
  }
EXIT STAGE: tester.workflow.report → terminal_success

ENTER STAGE: tester.workflow.terminal_success
EXIT STAGE: tester.workflow.terminal_success (workflow complete)

---

Quick Reference

State	Type	Exit Code → Target
load_tests	shell	0 → run_tests, else → terminal_error
run_tests	claude	default → verification_setup
verification_setup	shell	0 → verify, else → terminal_error
verify	claude	default → merge
merge	shell	0 → report, else → terminal_error
report	shell	default → terminal_success

Test Mode Reference

Mode	Description	Use Case
all	Run all tests in definition	Full test suite
single	Run one specific test	After implementing a feature
ticket-completion	Run only completion tests	Final verification before completion