onejaejae/bugfix

/bugfix Skill

Root cause based one-shot bug fix. Diagnose → generate spec.json → delegate to /execute for fix, verification, and commit.

When to Use (Differentiator)

| Skill | Purpose | When to pick | |-------|---------|-------------| | /bugfix | Root-cause diagnosis + spec generation + /execute delegation | Bug report — need investigation before fix | | /ralph | DoD-based iterative loop with independent verification | Known fix needed, want verified-done loop | | /execute | Spec-driven orchestrator from existing spec.json | Already have a spec.json, no diagnosis needed | | /scope | Fast parallel change-scope analyzer | Need to understand blast radius before fixing |

The Iron Law

NO FIXES WITHOUT ROOT CAUSE INVESTIGATION FIRST
No completion claims without evidence
Must stop after 3 failed attempts

Architecture

/bugfix "error description"

Phase 1: DIAGNOSE ─────────────────────────────────
  debugger + verification-planner (always parallel)
  + gap-analyzer (COMPLEX only)
  → User confirmation

Phase 2: SPEC GENERATION ──────────────────────────
  Diagnosis results → spec.json (hoyeon-cli spec init + merge)

Phase 3: EXECUTE ──────────────────────────────────
  Skill("execute", args=spec_path)
  → Success: Phase 5
  → HALT: Phase 4

Phase 4: RESULT HANDLING (if HALT) ────────────────
  SIMPLE: retry (max 3) with stagnation detection → Phase 3
  COMPLEX: escalate immediately (execute already retried)
  Circuit breaker → .dev/debug/{slug}.md → suggest /specify

Phase 5: CLEANUP & REPORT ─────────────────────────
  Save .dev/debug/{slug}.md → final summary

Adaptive Mode

Mode is never asked from the user. Auto-routes based on debugger's Severity assessment:

| Severity | spec depth | Phase 1 | Phase 3 (execute) | Retry | |----------|-----------|---------|-------------------|-------| | SIMPLE | quick | debugger + verification-planner | worker → commit → final verify | bugfix-managed (max 3) | | COMPLEX | standard | + gap-analyzer | worker → verify → commit → code review → requirements | execute-internal (max 2), then escalate |

---

Phase 1: DIAGNOSE

Step 1.1: Parse Input

Extract from user input:

• Bug description: error message, symptoms, reproduction steps
• Error output: stack trace, test failure logs (if available)
• Context: related files, recent changes (if available)

Initialize Debug State:

SESSION_ID = [from hook — $CLAUDE_SESSION_ID]
slug = convert bug description to kebab-case (e.g. "null-pointer-in-auth")
DEBUG_STATE = "$HOME/.hoyeon/$SESSION_ID/debug-state.md"
hoyeon-cli session set --sid $SESSION_ID --skill bugfix --debug "$DEBUG_STATE"

Write(DEBUG_STATE):
# Debug: {bug description}
status: investigating
severity: pending
attempt: 0
slug: {slug}

## Symptoms (IMMUTABLE after Phase 1)
- expected: {from user input}
- actual: {from user input}
- error: {from user input}

## Diagnosis
root_cause: pending
spec_path: pending

## Attempts

Step 1.2: Parallel Investigation

Always dispatch 2 agents in parallel:

Task(debugger):
  "Bug Description: {user input}
   Error Output: {error logs, if available}
   Context: {related files/recent changes, if available}

   Investigate this bug following your Investigation Protocol.
   Classify Bug Type, trace backward to root cause, assess Severity."

Task(verification-planner):
  "User's Goal: Fix the bug described below
   Current Understanding: {user input}
   Work Breakdown:
   - Reproduce bug with test
   - Apply minimal fix at root cause
   - Verify fix resolves the issue

   Focus on Auto items only (what commands prove the fix works).
   Keep it minimal — this is a bug fix, not a feature.

   Note: /bugfix uses Tier 1-3 (Auto items) only. Do not inline VERIFICATION.md.
   Tier 4 (sandbox items) are not needed. Mark sandbox section as 'bugfix mode — Tier 1-3 only'."

After receiving debugger results, update debug-state.md:

Update DEBUG_STATE:
  severity: {SIMPLE/COMPLEX}

  ## Diagnosis section:
  root_cause: {debugger's Root Cause — 1 line}
  bug_type: {classification}
  proposed_fix: {proposed fix — 1 line}

Step 1.3: Evaluate Severity & Conditional Gap Analysis

Check Severity from debugger results:

• SIMPLE → Proceed directly to Step 1.4
• COMPLEX → Run gap-analyzer, then Step 1.4:

Task(gap-analyzer):
  "User's Goal: Fix the bug below
   Current Understanding: {debugger's full Bug Analysis Report}
   Intent Type: Bug Fix

   Focus on:
   - Whether root cause vs symptom distinction is correct
   - Whether proposed fix could break other areas
   - Whether similar bugs exist with the same pattern"

Step 1.4: User Confirmation

Present debugger results summary for user confirmation:

AskUserQuestion:
  header: "Root Cause"
  question: "Is the debugger's Root Cause analysis correct?"

  Display:
  - Bug Type: [classification]
  - Root Cause: [file:line + 1-line description]
  - Proposed Fix: [change description — 1 line]
  - Severity: [SIMPLE/COMPLEX]
  - Verification: [verification commands from verification-planner]
  - Assumptions: [debugger's Assumptions section]
  [COMPLEX only] Key warnings from Gap Analysis

  options:
  - "Correct, proceed" → Phase 2
  - "Root cause is different" → Re-run Step 1.2 with user's additional info
  - "Not sure" → Suggest "/discuss to explore first", then exit

---

Phase 2: SPEC GENERATION

Convert diagnosis results into spec.json v5 format. spec.json is the standard format consumed by /execute, and serves as escalation context for /specify on failure.

Step 2.1: Initialize

SPEC_DIR = "$HOME/.hoyeon/$SESSION_ID"
SPEC_PATH = "$SPEC_DIR/spec.json"

depth = "quick" for SIMPLE, "standard" for COMPLEX

hoyeon-cli spec init fix-{slug} \
  --goal "Fix: {bug description}" \
  --type dev \
  --depth {depth} \
  --interaction autopilot \
  ${SPEC_PATH}

Step 2.2: Merge diagnosis into spec

Use hoyeon-cli spec merge to populate the spec from diagnosis results. Single merge call.

What to include:

• meta: non_goals (no refactoring, no unrelated features), deliverables (debug report path)
• context: request (original bug description), research (debugger analysis summary), assumptions (from debugger), decisions (root cause location + rationale)
• tasks: Single task (T1) with:
  • action: debugger's proposed fix
  • file_scope: affected files from debugger
  • steps: write regression test (RED) → apply minimal fix (GREEN) → verify
  • must_not_do: minimal diff (<5%), no refactoring, no unrelated changes, no git commands, fix root cause not symptom
  • acceptance_criteria.scenarios: list of scenario IDs from requirements[].scenarios[].id (verification-planner's Auto items map to machine scenarios)
  • acceptance_criteria.checks: automated static/build/lint checks if applicable
  • If debugger found similar issues: add T2 (depends_on: [T1]) to fix those locations — T2 must include all required task fields: type: "work", status: "pending", must_not_do, and acceptance_criteria
• constraints: minimal diff rule, root cause targeting rule (both verified_by: agent)
• requirements: Generate from debugger diagnosis. Each requirement describes a behavior that was broken:

  {
    "id": "R1",
    "priority": 1,
    "behavior": "{what should work — from debugger's expected behavior}",
    "scenarios": [{
      "id": "R1-S1",
      "given": "{precondition from debugger's reproduction steps}",
      "when": "{trigger action that caused the bug}",
      "then": "{expected outcome after fix}",
      "verified_by": "machine",
      "verify": {
        "type": "command",
        "run": "{test command from verification-planner's Auto items}",
        "expect": {"exit_code": 0}
      }
    }]
  }
  

  • Convert debugger's reproduction steps → Given/When/Then
  • Use verification-planner's Auto items as verify.run commands
  • If debugger identified edge cases, add additional scenarios
  • This enables Final Verify to check requirements scenarios, preventing regression

Step 2.3: Validate & Register

hoyeon-cli spec validate ${SPEC_PATH}
hoyeon-cli session set --sid $SESSION_ID --spec "$SPEC_PATH"

If validation fails, fix the JSON and retry once.

Update debug-state.md with spec_path: ${SPEC_PATH}.

---

Phase 3: EXECUTE

Hand off spec.json to /execute. Execute routes by meta.type: dev and follows the dev.md pipeline.

Skill("execute", args="${SPEC_PATH}")

What execute handles:

• Worker dispatch (self-read pattern)
• Per-task verify (standard only)
• Per-task commit (git-master)
• Retry/Adaptation (standard only, max 2)
• Code Review (standard only)
• Final Verify (all modes — checks goal, constraints, AC, requirements, deliverables)
• Final report

Result judgment:

IF execute completed successfully (all tasks done, report output):
  → Phase 5

IF execute HALTED:
  → Phase 4

---

Phase 4: RESULT HANDLING

When execute HALTs. Handling differs by severity.

Step 4.1: Read Failure Context

# Extract failure reason from execute's HALT output
# or read from context dir's audit.md, issues.md
CONTEXT_DIR = ".dev/specs/fix-{slug}/context"
failure_reason = {execute HALT output or last triage result from audit.md}

Step 4.2: Route by Severity

IF severity == "COMPLEX":
  # Execute standard mode already retried internally (max 2)
  # No further retries — go straight to Circuit Breaker
  → Step 4.5 (Circuit Breaker)

IF severity == "SIMPLE":
  → Step 4.3 (Retry)

Step 4.3: Retry (SIMPLE only)

# Read current attempt from debug-state.md
attempt = debug_state.attempt + 1
MAX_ATTEMPTS = 3

IF attempt >= MAX_ATTEMPTS:
  → Step 4.5 (Circuit Breaker)

Stagnation Detection (attempt >= 2):

# Compare with previous attempt failure info
previous = debug_state.Attempts[-1]
current_reason = failure_reason

SPINNING:    same file/component fails consecutively
OSCILLATION: A fails → B fails → A fails (circular)
NO_PROGRESS: different failures each time, previous fixes cause regressions

Pattern-specific retry_hint:
  SPINNING    → "Different root cause likely. Consider: previous root cause
                 was wrong — trace further back from the symptom."
  OSCILLATION → "Circular dependency. Fix both sides simultaneously."
  NO_PROGRESS → "Fundamental misunderstanding. Re-read error output.
                 Consider: multiple independent bugs? Missing dependency?"
  (no pattern) → "Different approach needed. Do NOT repeat previous attempt."

Step 4.4: Update Spec & Re-execute

# 1. Record attempt in debug-state.md
Append to DEBUG_STATE ## Attempts section:
  ### Attempt {attempt}
  result: FAIL
  reason: {failure_reason}
  pattern: {detected pattern or "none"}
  hint: {retry_hint}

# 2. Update attempt counter
Update DEBUG_STATE: attempt: {attempt}

# 3. Add failure context to spec.json
hoyeon-cli spec merge ${SPEC_PATH} --json '{
  "context": {
    "known_gaps": [
      {"gap": "Attempt {attempt} failed: {failure_reason}", "severity": "high", "mitigation": "{retry_hint}"}
    ]
  }
}'

# 4. Reset task status
hoyeon-cli spec task T1 --status pending ${SPEC_PATH}
{IF T2 exists AND T2 not done:}
hoyeon-cli spec task T2 --status pending ${SPEC_PATH}

# 5. Re-invoke execute
→ Phase 3

Execute handles resume naturally:

• Skips done tasks
• Context files (learnings.md, issues.md) retain previous failure info for the new worker
• known_gaps carry failure context and retry_hint for the worker

Step 4.5: Circuit Breaker

Max attempts exceeded or COMPLEX mode HALT. Present escalation options to user.

First, save attempt records:

Bash: mkdir -p .dev/debug

Write to .dev/debug/{slug}.md:
  # Bugfix Report: {description}
  Date: {timestamp}
  Status: ESCALATED
  Severity: {SIMPLE/COMPLEX}
  Attempts: {attempt count}
  Spec: {SPEC_PATH}

  ## Debugger Analysis
  {debugger's full Bug Analysis Report}

  ## Attempt History
  {full ## Attempts section from debug-state.md}

  ## Assessment
  {SIMPLE: "{attempt} attempts failed. Likely not a simple bug."}
  {COMPLEX: "Execute standard mode failed including internal retries. Architecture-level issue."}

Update DEBUG_STATE:
  status: escalated

AskUserQuestion:
  header: "Circuit Breaker"
  question: "Fix attempts have failed. This may be an architecture-level issue."
  options:
  - "Switch to /specify (full planning)"
    → "spec.json and debug report are available:
       Spec: {SPEC_PATH}
       Report: .dev/debug/{slug}.md
       /specify can reference this context for deeper analysis."
  - "Try once more"
    → attempt += 1, go to Phase 3 (no circuit breaker reset)
  - "Stop"

---

Phase 5: CLEANUP & REPORT

After execute completes successfully.

Step 5.1: Save Debug Report

Bash: mkdir -p .dev/debug

Write to .dev/debug/{slug}.md:
  # Bugfix Report: {description}
  Date: {timestamp}
  Status: RESOLVED
  Severity: {SIMPLE/COMPLEX}
  Attempts: {attempt count + 1}
  Spec: {SPEC_PATH}

  ## Root Cause
  {debugger's Root Cause analysis}

  ## Fix
  {spec.json T1.action + result summary}

  ## Verification
  {verification results from execute's final report}

Update DEBUG_STATE:
  status: resolved

Step 5.2: Final Summary

print("""
## Bugfix Complete

**Bug**: {description}
**Root Cause**: {file:line — 1-line description}
**Severity**: {SIMPLE/COMPLEX}
**Attempts**: {count}
**Spec**: {SPEC_PATH}
**Report**: .dev/debug/{slug}.md
""")

---

Escalation Path

/bugfix (diagnose + spec.json + execute)
   ↓ circuit breaker (SIMPLE: 3 failures, COMPLEX: execute HALT)
   ↓ spec.json + .dev/debug/{slug}.md saved
/specify (spec.json enrichment, leveraging existing diagnosis context)
   ↓
/execute (enriched spec execution)

Since spec.json is the standard format, /specify can read and enrich the existing spec on escalation. All diagnosis context (context.research, context.assumptions, context.known_gaps) is preserved.

---

Agent Summary

| Phase | Agent | Status | Condition | Role | |-------|-------|--------|-----------|------| | 1 | debugger | existing | always | Root cause analysis, Bug Type classification, Severity assessment | | 1 | verification-planner | existing | always | Generate Auto items list (verification commands) | | 1 | gap-analyzer | existing | COMPLEX only | Check for missed factors, risk assessment | | 3 | /execute (Skill) | existing | always | spec.json-based execution (worker, verify, commit, review) |

Phase 2 (SPEC GENERATION) and Phase 4 (RESULT HANDLING) are handled directly by bugfix without agents (hoyeon-cli calls + judgment logic).

---

Design Principles

This skill combines core patterns from 3 proven open-source projects:

| Principle | Source | Application | |-----------|--------|-------------| | Root cause before fix | superpowers (systematic-debugging) | Entire Phase 1 | | Backward call stack tracing | superpowers (root-cause-tracing) | debugger's Step 3 | | Defense-in-depth after fix | superpowers (defense-in-depth) | Optional worker application | | Anti-pattern rationalizations | superpowers (common rationalizations) | debugger's checklist | | Bug Type → Tool routing | oh-my-opencode (Metis intent classification) | debugger's tool table | | Adaptive severity | oh-my-opencode (Momus "80% is good enough") | SIMPLE/COMPLEX auto-routing | | Minimal diff (<5%) | oh-my-claudecode (executor/build-fixer) | spec constraint C1 | | Circuit breaker (3 attempts) | oh-my-claudecode (debugger) + superpowers | Phase 4 | | spec.json as universal format | internal (specify/execute unification) | Phase 2 | | Execute reuse | internal (single execution engine) | Phase 3 |