mirror of
https://github.com/davila7/claude-code-templates.git
synced 2026-09-19 01:30:23 +08:00
Daniel Avila
7f890b1b3b
* feat: add Loops component type with referenced components Introduce "loops" as a new first-class component type — autonomous agentic workflows (goal + interval + stop condition) inspired by Loop Engineering and the Forward Future Loop Library. A loop references other components (agents, skills, hooks, commands) and the dashboard renders those references as clickable cards on the loop detail page. - Add 10 starter loops under cli-tool/components/loops/ (engineering, evaluation, operations) - Catalog generator: scan loops + write per-component content files - Dashboard: new type in icons/types/data, /loops page, and a "Referenced Components" section on the loop detail page - CLI: --loop flag installs the loop to .claude/loops/ and auto-installs its referenced components - Docs: CLAUDE.md and cli-tool rule updated Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01T9ynMiCHt8JwdMd8KjCkES * feat: add 8 more loops (goal/loop/schedule patterns) Expand the loops catalog to 18, drawing ideas from community "Loop Engineering" patterns. Adds loops across the three trigger verbs (/goal, /loop, /schedule) and emphasizes the verifier-in-the-loop and budget / anti-spin guardrails: - engineering/five-minute-maintainer-loop - engineering/build-test-fix-loop - engineering/anti-spin-build-loop - engineering/completion-contract-loop (goal) - engineering/goal-refiner-loop (goal) - engineering/adversarial-review-loop (cross-model) - operations/overnight-pr-routine-loop (schedule) - evaluation/human-approval-loop (human-in-the-loop) All referenced components verified to exist; catalog regenerated. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01T9ynMiCHt8JwdMd8KjCkES --------- Co-authored-by: Claude <noreply@anthropic.com>
43 lines
2.1 KiB
Markdown
43 lines
2.1 KiB
Markdown
---
|
||
name: builder-reviewer-loop
|
||
description: Passes code between a builder and an independent reviewer subagent until the review has no blocking findings and tests pass.
|
||
category: engineering
|
||
interval: 15m
|
||
stop-condition: The reviewer reports zero blocking findings and the full test suite passes.
|
||
components: [agent:expert-advisors/architect-review, agent:development-team/test-runner, command:git-workflow/pr-review]
|
||
tags: [code-review, subagents, quality, loop]
|
||
---
|
||
|
||
# Builder–Reviewer Loop
|
||
|
||
> **Loop Engineering** — separate the *builder* from the *reviewer* so the agent doesn't grade its own work. Code bounces between them until it genuinely passes.
|
||
|
||
## 🎯 Goal
|
||
Implement a change, then have an **independent reviewer subagent** critique it; iterate on the findings until the review is clean and tests are green.
|
||
|
||
## ⏱️ Schedule
|
||
Suggested interval: `15m` (or run until done on a single feature).
|
||
|
||
## ▶️ Run it
|
||
```
|
||
/loop 15m "Implement the next slice of the feature. Then hand it to an independent reviewer subagent for an adversarial review. Address every blocking finding and re-run tests. Continue until the reviewer has no blocking findings and the suite passes."
|
||
```
|
||
|
||
## 🔁 Iteration steps
|
||
1. **Build** — implement the next slice of work.
|
||
2. **Review** — the `architect-review` agent inspects the diff for correctness, design and risk.
|
||
3. **Plan** — collect blocking findings into a fix list.
|
||
4. **Act** — address each finding; re-run the suite with `test-runner`.
|
||
5. **Observe** — re-review with `/pr-review`; if blocking findings remain, loop again.
|
||
|
||
## 🛑 Stopping condition
|
||
Reviewer reports zero blocking findings **and** all tests pass.
|
||
|
||
## 🧩 Referenced components
|
||
- `agent:expert-advisors/architect-review` — the independent reviewer (no self-grading).
|
||
- `agent:development-team/test-runner` — runs the suite each pass.
|
||
- `command:git-workflow/pr-review` — structured review checklist.
|
||
|
||
## 💡 Example
|
||
The builder ships a new caching layer; the reviewer flags a race condition and a missing eviction test. The loop fixes both, re-reviews, and exits clean.
|