Home / All Skills / Workflow & Core Tools / loop-me
English (EN)简体中文 (ZH)日本語 (JA)
Workflow & Core Tools · mattpocock/skills

loop-me

Grill me about specs for the workflows I want to build, within this workspace.

L3 Networked · Declare domain & key scopeV1 Commit-PinnedNetwork AccessWrite FileDual-pass Audit
Author mattpocockGitHub mattpocock/skills ↗Stars 257659License MIT（GitHub API spdx MIT；仓库根 LICENSE: Copyright (c) 2026 Matt Pocock）Commit 3cca18b368
Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Stateful encapsulation powered by grilling: Directly runs a /grilling session dedicated solely to producing workflow specs, reusing its relentless discipline, multi-question rounds, and recommended answers per question.

skills/in-progress/loop-me/SKILL.md
Run a stateful `/grilling` session whose only output is **workflow** specs. Use the grilling discipline (relentless, a round of questions at a time, a recommended answer attached to each) aimed at the vocabulary and goal below. Create, edit, and delete specs as the grilling resolves things.
What is happening here: loop-me does not invent a new questioning mechanism; instead, it wraps grilling with domain vocabulary (loop/workflow/trigger/checkpoint/push right/brief) and output constraints (workflow specs), authorizing direct workspace file writes during the grilling process.

Loop lens to identify delegable targets: Treats recurring patterns in a user's life (career/week/morning/single repetitive activity) as loops, where loop predictability equals justification for delegation, actively suggesting loops the user may not have noticed.

skills/in-progress/loop-me/SKILL.md
A **loop** is a recurring pattern in the user's life: their career, their week, their morning, a single repeated activity. Picturing a life as loops within loops reveals how predictable its activities really are, which is what makes them worth **delegating**. Use the lens to find loops worth specifying, and propose ones the user hasn't noticed.
Workflow is the concrete spec entity of a loop: 'A **workflow** is the spec of one loop, made real. You run a workflow on a loop: the loop is its running instantiation.'

Principle of vocabulary restraint: The shared language is used only when the workflow demands it and is never treated as a checklist; zero structural enforcement (AI/checkpoint/schedule are never mandatory unless proven necessary by grilling).

skills/in-progress/loop-me/SKILL.md
A shared language, reached for only when a workflow calls for it: never a checklist. **Mandate nothing structural**: a workflow needs no AI, no checkpoint, and no schedule unless the grilling shows it does.
This is an explicit rejection of 'all workflows must be automated/AI-driven'—purely manual loops are fully within scope.

Two core design principles: push right (defer human checkpoints as late as possible: ask once, ask late, prepare everything) and brief (present decision-ready briefs to the user at checkpoints rather than raw outputs).

skills/in-progress/loop-me/SKILL.md
**Push right**: defer the checkpoint as far as it will go. Do maximal work before involving the human, so they are asked once, late, with everything prepared. **Brief**: what a checkpoint presents, a tight, decision-ready summary (what was produced, why, and a link down to the asset itself), never the raw output. The user reads a brief, not a draft. Speed of review is imperative.
This defines the behavioral aesthetics of the generated workflows: maximizing unattended runtime while keeping human intervention points minimal and refined.

Completion criteria = implementable without questions: Grilling continues until 'an implementer agent can build it to spec without asking a single question'; as long as questions remain, it is incomplete.

skills/in-progress/loop-me/SKILL.md
A workflow spec is done when an implementer agent could build it without asking a single question. Grill until then; nothing is done while a question remains.
Translating grilling's 'empty frontier = complete' to this domain: all unknown facets of the specification are cleared to zero.

Dual-file workspace as cross-session state: workflows/*.md serves as one spec per loop (source of truth), while NOTES.md stores raw observations and terminology of the user's world; when NOTES is empty or thin, interview the user's world before starting specification.

skills/in-progress/loop-me/SKILL.md
The workspace - `workflows/*.md`: one spec per workflow. - `NOTES.md`: raw notes on the user's world, the tools they use, the channels they process, and their own terminology for both. When it is empty or thin, interview them about their world before specifying anything. Sharpen fuzzy terms into canonical ones as they surface, and record them here.
The in-progress README describes it as 'over multiple sessions, using the current directory as a stateful workspace'—the state consists of these two files.

2Core Capabilities

01Identify and propose unnoticed delegable loops for the user (recurring life patterns → workflow candidates)
02Multi-session stateful grilling: Maintain workflows/*.md and NOTES.md across sessions
03Generate specifications according to loop vocabulary (trigger/checkpoint/push right/brief) with zero structural enforcement
04Front-load world modeling: Interview the user's world, tools, and terminology when NOTES is empty before specifying
05Use 'implementable by a builder without asking questions' as the completion gate

4Recommended Governance & Evidence

Suggested Controls for Safe Operation
For Agent Runtime
Declare allowed domains and key scope
For Administrators
Least privilege keys, retain request logs
Evidence & Triggers (File Paths & Excerpts)
  • 本地工作目录: File I/O and artifact output: create/edit/delete workflows/*.md in the current workspace (explicitly authorized); create and append to NOTES.md
  • 外部服务: Relies on host network tools or external APIs for retrieval
Audit Notes: Behavior = grills the user + creates, edits, and deletes workflows/*.md and NOTES.md in the current workspace. The skill explicitly authorizes: 'Create, edit, and delete specs as the grilling resolves things'. Because file deletion is a destructive local operation (scoped strictly to the workflows/*.md it manages, not arbitrary paths or total disk wipes), it is classified one tier higher as Yellow according to unified tiers. No network exfiltration, no credential reading, and no script execution. Recommendations: launch inside a dedicated working directory, confirm the directory path beforehand, and review generated specs before handing them off for implementation.

5Second-Pass Verification

  • [ok] 复用 /grilling 引擎声明 — SKILL.md 首句 'Run a stateful `/grilling` session whose only output is **workflow** specs' 存在；grilling 为 model-invoked 原语，调用方向合规。
  • [ok] 文件操作范围（workflows/*.md、NOTES.md create/edit/delete） — 原文 'Create, edit, and delete specs as the grilling resolves things.' + The workspace 段；无任何工作区外路径/绝对路径/删除指令。
  • [ok] user-invoked 双清单 — disable-model-invocation: true + policy.allow_implicit_invocation: false 均在；argument-hint 亦在 frontmatter。
  • [ok] loop lens / 词汇表 / push right / brief 概念齐全 — 逐段存在（A **loop** is…、Push right、Brief、Trigger、Checkpoint）。
  • [ok] 完成门禁（implementer 无问可建） — 原文 'A workflow spec is done when an implementer agent could build it without asking a single question.' 存在。
  • [ok] 外部依赖为空 — token 扫描零命中；无 CLI/API/网络/包引用。grilling 是仓库内 skill 引用非外部依赖。
  • [ok] 元数据 + commit — GitHub API MIT/257659 stars/2026-09-04T08:45:43Z；本地 HEAD == 3cca18b。
  • [ok] beta 状态标注 — in-progress/README.md 将 loop-me 列于 in-progress 并声明 excluded from plugin、'can change or disappear without warning'，与 verdict 风险描述一致。

6Verdict & Scenarios

  • Transforms 'grilling' into a workflow specification engine: directly reuses grilling discipline without reinventing questioning mechanisms.
  • Clear design philosophy: push right + brief makes generated workflows inherently low-interruption and brief-oriented; 'Speed of review is imperative'.
  • Anti-AI dogma: zero structural enforcement, explicitly permitting workflows with no AI, no checkpoints, or no schedules.
  • Front-loaded world modeling: interviews the user's world and terminology when NOTES is empty, ensuring specs are not built on misunderstandings.
  • Cross-session statefulness: a dual-file workspace enables multi-round interviews to accumulate into implementable specifications.
    • Best for: Suitable for users who want to systematically delegate repetitive work and life activities to agents and are willing to undergo multi-round grilling interviews; especially ideal for accumulating a 'personal workflow library' across sessions in a dedicated workspace directory, producing outputs ready for immediate handover to an implementer agent.
    Not for: Not suitable for one-off quick Q&A (it is a stateful, long session that writes files); not suitable for users who do not want skills modifying workspace files; not suitable for team-shared repositories (spec writing carries a strong personal perspective and is in an unstable beta); users seeking generic engineering artifacts (spec/ADR) should use to-spec/grill-with-docs instead.
    Install Agent-Ready
    ① Verified Mirror 2026-09-09
    Option A · Direct DownloadDownload loop-me.tar.gz
    sha256: 7815f818e390fcec…
    Option B · Copy Guide for Agent
    install.json
    Agents automatically download and verify mirrors from this JSON guide.
    ② Upstream GitHub
    GitHub Repo ↗
    Pinned to commit 3cca18b368.