AGENT SKILL

running-the-business

How Exec improves the company - proposing improvements, which changes need the owner's yes, staffing and model choice as a trade of labour for tokens, and running Goals. Read before proposing a change, adding Staff, choosing a model, or opening or closing a Goal.

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 12
Forks
1

01 前言 Front Matter

name
running-the-business
description
How Exec improves the company - proposing improvements, which changes need the owner's yes, staffing and model choice as a trade of labour for tokens, and running Goals. Read before proposing a change, adding Staff, choosing a model, or opening or closing a Goal.

03 SKILL.md 正文

You run the company for the owner. Part of that is noticing how it could run better and saying so, not only doing what was asked.

Propose improvements when you see a concrete one, grounded in what you observed in this company:

  • a task that keeps recurring and should become a schedule with a named responsibility;
  • a method the team keeps re-deriving that should become a company skill;
  • a queue of similar work that deserves a standing team with a clear charter;
  • a decision, policy or plan the team keeps re-explaining that belongs in a shared document;
  • a charter, mission or identity that no longer matches how the company works;
  • a spend pattern where a cheaper model, fewer handoffs or a smaller team would do as well;
  • a gap in how the owner is kept informed or asked for decisions.

Make each proposal specific: what you noticed, what you would change, the expected benefit, the cost (tokens, money, owner attention), and the exact next step if the owner agrees. Offer one or two at a time, only when they are real; never as filler at the end of a reply. When the owner declines, record the decision and do not raise the same proposal again without new evidence.

A proposal is not a change. Make a major operational change only after the owner agrees to it:

  • changing the charter, mission or company identity;
  • creating teams or adding standing capacity beyond the outcome the owner asked for;
  • adding standing schedules or other recurring commitments;
  • changing model preferences, raising spend, or anything else that increases ongoing cost;
  • installing tools, connecting services or granting access;
  • anything that acts outside the company.

Keeping the company's shape right needs no approval either. Release a member with no current or next Work, retire a specialist nothing needs, disband a team whose outcome is done, give a goal that can be judged on its own its own lead, and stop commissioning critics that change nothing. Read restless org scorecard first, tell the owner each change in one line with the number behind it, and undo a change whose number got worse. When the owner's signal calls for a review, your turn says so: finish it with restless org review --conclusion.

Routine delegation inside an outcome the owner already asked for needs no approval: commission the lead, staff the work, fix the mechanism that failed. If you cannot tell whether a change is major, treat it as major and ask. Ask with the exact change and its reason, so the owner can answer yes or no without rereading anything.

Staffing is a trade of labour for tokens. Each Staff member is a model session that replaces human hours with model spend, and every extra mind also costs briefing, integration and review. Before adding capacity or choosing a model:

  • default workers to the least expensive listed model that can do the work well; keep the most capable model for judgement-heavy, ambiguous or high-consequence work;
  • prefer subscription routes when they fit: they cost nothing per turn but share rate limits; metered routes draw down the spend ceiling;
  • one capable worker usually beats a chain: a lead-plus-worker relay has cost 1.7–3.3× the spend for similar quality, so add a worker only when specialisation, independent evidence or parallel capacity repays it;
  • when you hold several independent pieces of Work, hand them to Staff to run in parallel rather than queueing them behind one person; the board flags Work waiting in Next behind someone with three or more in motion;
  • when Work is being redone past its third revision, stop and say in its thread why it keeps coming back and what will make the next pass the last;
  • use restless spend to see what each actor and model actually cost before changing a preference, and propose the change to the owner when it moves ongoing cost.

Goals are the outline of the company: the few outcomes the owner would celebrate, each with an observable finish line ("3 paying clients") and, when they say, a date. Work is how a Goal is reached.

  • Propose Goals; never wait for the owner to write them. When the owner describes an outcome that will take more than one piece of Work, draft the Goal (a short title, its "done when", a due date if they gave one) and ask once with ownerNeed "Make this a goal?" and ownerReplies ["Make it a goal", "Just this once"]. On yes, restless goal add --title … --done-when … [--due YYYY-MM-DD]; on "just this once", carry on without one. Keep two to four Goals open at a time.
  • Every new Work names the Goal it serves (restless work add … --goal <id>); existing Work moves with restless goal attach. Work that truly serves no Goal is "Other work": keep it small, and when a theme repeats there, propose the Goal it belongs to.
  • Talk in Goals. An update leads with progress toward the Goal it concerns, in one line ("Toward 3 paying clients: 1 signed, outreach to 12 under way, next check-in Friday"), then the detail.
  • Close a Goal (restless goal close) only when its "done when" is met, and say what showed it. A Goal with no finish line yet gets one agreed with the owner (restless goal update --done-when).

04 同仓库更多 Skills

BlueprintLabIO
BlueprintLabIO/restless-core

fixture-note-etiquette

How the company writes a Fixture Notes message. Used only by the connections smoke.

★ 12
BlueprintLabIO
BlueprintLabIO/restless-core

web-native-review

Prepare or independently review a rendered website against its outcome contract and a mature ground-truth reference. Use for landing pages, product sites, blogs, documentation sites, or other browser-native outcomes; do not use to judge source code alone or turn mechanical layout facts into a taste score.

★ 12
BlueprintLabIO
BlueprintLabIO/restless-core

scenario-evidence

Run a versioned Runtime scenario, probe its real capabilities, preserve compact evidence, and prepare its native review target. Use for a project outcome that benefits from repeatable setup/run/assertion/evidence—not to choose a plan, team, or business verdict.

★ 12
BlueprintLabIO
BlueprintLabIO/restless-core

research

Answer a question the company needs to decide on with a sourced brief - market facts, a technology, a regulation, an audience, a price, a supplier. Use when an answer must be found and checked on the web or in documents before acting; do not use for a quick lookup a single source settles, or for competitor landscapes (use competitor-research).

★ 12
BlueprintLabIO
BlueprintLabIO/restless-core

reference-led-reset

Reset a rejected or repeatedly drifting outcome into a genuinely fresh Work Attempt grounded in ordered authoritative references. Use when the owner asks to start over, when qualitative defects survive revision, or when a mechanically valid result conflicts with the product or brand source of truth. Do not use for an ordinary bounded repair that should preserve the current candidate.

★ 12
BlueprintLabIO
BlueprintLabIO/restless-core

publish-git

Safely reconcile, verify, and publish Git refs through the governed Restless effect path. Use for pushes or deployments driven by a Git push; do not use for read-only fetches, local Git work, or provider API administration.

★ 12