/implement-inplace

User request: $ARGUMENTS

Autonomously execute implementation in-place. Supports plan files, inline tasks, or interactive mode.

Fully autonomous: No pauses except truly blocking issues.

Workflow

Phase 1: Resolve Input & Setup

Priority order:

1. File path (ends in .md or starts with /) → use plan file, optionally with --spec <path>
2. Inline task (any other text) → create ad-hoc single chunk:

   ## 1. [Task summary - first 50 chars or natural break]
   - Depends on: -
   - Tasks: [user's description]
   - Files: (discover during implementation)
   - Acceptance criteria: gates pass
   

3. Empty → search /tmp/plan-*.md (most recent); if none found, ask user what they want to implement

For plan files, parse each chunk (## N. [Name] headers):

• Dependencies (Depends on: field, - = none)
• Files to modify/create with descriptions
• Context files (paths, optional line ranges)
• Implementation tasks (bullet list)
• Key functions/types

Build dependency graph: No-dependency chunks first, then topological order.

Create flat todo list (Memento pattern—granular progress, resumable):

[ ] Read context for [Chunk]
[ ] [Task 1]...[ ] [Task N]
[ ] Run gates for [Chunk]

All todos created at once via TodoWrite, status pending.

Phase 2: Execute Chunks

CRITICAL: Execute continuously without pauses.

Per chunk:

1. Read context files (related + files-to-modify), respect line ranges
2. Implement each task, marking in_progress→completed immediately
3. Run quality gates (Phase 3)
4. Track created/modified files for summary

Spec file (--spec): Read before implementation for requirements/acceptance criteria.

Phase 3: Auto-Fix Gates

Run gates (typecheck→tests→lint). Detect commands from CLAUDE.md first, then package.json/pyproject.toml.

On failure—iterate:

1. Analyze: parse errors, identify files/lines, understand root cause
2. Fix cleanly (no workarounds)
3. Re-run gates
4. Repeat until pass, track attempts to avoid loops

Pause ONLY when: 3+ failed iterations, need external info, requires out-of-scope breaking changes, fundamental conflict.

Report: what tried, why failed, what's needed.

Phase 4: Completion

## Implementation Complete
Chunks: N | Todos: M | Created: [list] | Modified: [list]
### Notes: [warnings/assumptions/follow-ups]
Run `/review` for quality verification.

Edge Cases

Case	Action
Invalid plan	Error with path + expected structure
Missing context file	Warn, continue (non-blocking)
Chunk fails completely	Leave todos pending, skip dependents, continue independents, report in summary
Inline task provided	Create ad-hoc single chunk, proceed normally
No input + no recent plan	Ask user what they want to implement
Interrupted	Todos reflect exact progress (Memento), manual cleanup needed for re-run

Principles

• Autonomous: No prompts/pauses/approval needed except blocking issues
• Memento todos: One todo per action, granular visibility, resumable
• Persistent auto-fix: Iterate until gates pass, escalate only when stuck
• Dependency order: Execute in order, skip failed chunk's dependents
• Gates non-negotiable: Fix now, not later

Gate Detection

Priority: CLAUDE.md → package.json scripts → Makefile → config detection

Fallback (if CLAUDE.md doesn't specify):

• TS/JS: tsconfig.json→tsc --noEmit, eslint.config.*→eslint ., jest/vitest.config.*→npm test
• Python: pyproject.toml→mypy/ruff check, pytest config→pytest