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>
2.4 KiB
name, description, category, interval, stop-condition, components, tags
| name | description | category | interval | stop-condition | components | tags | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| build-test-fix-loop | Builds the next item on the plan, then runs tests, typecheck and lint, feeding every failure back as the next instruction until the build is green. | engineering | 10m | The build is green — tests, typecheck and lint all pass — and the checker reports nothing left to fix. |
|
|
Build–Test–Fix Loop
Loop Engineering — the single most-demoed loop pattern: a builder writes code and a checker runs tests, typecheck and lint and reports exactly what broke. They pass work back and forth until it's clean. A one-shot agent ships its bugs; this loop catches them.
🎯 Goal
Implement the next item on the plan, then run the full check gate (tests + typecheck + lint), feed every failure back as the next instruction, and repeat until the build is green.
⏱️ Schedule
Suggested interval: 10m (or run until done on a feature).
▶️ Run it
/loop 10m "Build the next item on the plan, then run tests, typecheck and lint. Feed every failure back as the next instruction and fix it. Add a test for anything that lacked coverage. Stop when the build is green and the checker has nothing left to report."
🔁 Iteration steps
- Build — implement the next plan item.
- Check — the
test-runnerhook runs tests, typecheck and lint automatically. - Reason — collect every failure into a concrete fix list.
- Act — fix each; add tests with
/generate-testsfor uncovered paths. - Observe — re-run the gate; loop until fully green.
🛑 Stopping condition
Tests, typecheck and lint all pass and the checker reports nothing outstanding.
💰 Budget & guardrails
Cap iterations or spend so a stubborn failure can't spin forever — surface anything that fails the same way twice for human eyes.
🧩 Referenced components
agent:development-team/test-runner— runs and interprets the check gate.hook:testing/test-runner— fires the suite automatically each pass.command:testing/generate-tests— covers gaps the loop uncovers.
💡 Example
Implementing a new endpoint, the loop catches a type error, then a failing integration test, then a lint warning — fixing each in turn — and exits only when all three gates are green.