AGENT SKILL

progress-check

Give sparse, evidence-based progress updates during long-running work: observable milestones weighted by work and risk, an update only when something material changed, one rounded-down bar. Use while long work with several milestones runs between replies, never for a short answer or work with no intermediate state; and the moment the owner asks to change how or how often you report progress, whether more updates, total silence or an estimated clock, BEFORE any such preference is saved anywhere.

License 许可
MIT
最近更新
2026-10-21
Stars
★ 0
Forks
0

01 前言 Front Matter

name
progress-check
description
Give sparse, evidence-based progress updates during long-running work: observable milestones weighted by work and risk, an update only when something material changed, one rounded-down bar. Use while long work with several milestones runs between replies, never for a short answer or work with no intermediate state; and the moment the owner asks to change how or how often you report progress, whether more updates, total silence or an estimated clock, BEFORE any such preference is saved anywhere.

02 资源文件 4

03 SKILL.md 正文

Progress Check

progress-check is the habit for the moment long work runs between replies: a change with several milestones, a build of many slices, delegated agents running one after another. It never fires on a short answer or on work whose only honest states are started and finished, and its cadence is judgment, never spam. It also fires the moment the owner asks to change how or how often you report, whether more updates, silence or an estimated clock: follow the preference within the current task, keep reporting actionable blockers, and save nothing anywhere unless the owner asks for a lasting rule (Persist a preference, below). It writes no record of its own: progress stays in the conversation, the evidence behind it belongs to the change record's sections and the checks the change recorded, and an estimated clock is an unmeasured claim this habit never makes.

Greenline's paths and commands for this skill stand on its references page, references/greenline.md: read it when the work lands in the workspace.

Full method: METHODOLOGY.md. Chat shape: RESPONSE.md. Examples: EXAMPLES.md.

Choose the mode

  • Report progress: default for qualifying long-running work.
  • Persist a preference: only when the owner asks for a lasting rule. The greenline CLI installs this skill, so there is no install-globally mode here.

Report progress

  1. Define a small set of observable milestones before estimating a percentage. Weight by work and risk, not elapsed time. Where a change owns the work, the milestones are its acceptance items and every update quotes their evidence.
  2. Send one baseline update after scope is understood.
  3. Send a routine update only when turns of work have passed since the previous routine update or a step of the change has ended, and one of these is true: a milestone completed, verified progress rose by at least 10 percentage points, or the next action materially changed. Never estimate elapsed time to open the gate; an estimated clock is an unmeasured claim.
  4. Report a new blocker immediately when the user can act on it. Report completion immediately.
  5. Never repeat an unchanged bar or send a routine “still working” message.

The update is conversation and nothing else: milestone evidence belongs to the change record and its checks, and no parallel progress file or tracker is created. Keep each update to at most three short items: completed evidence, current work, and next step or blocker. End with exactly one overall bar styled as inline code on its own line:

Performance rollout [██████████████████░░] 90%

Use one pair of backticks for inline-code styling. Do not indent the line by four spaces and do not use a fenced code block; both create a separate block that interfaces may label or make copyable.

  • Replace the label with a short task-specific label.
  • Use 20 cells: █ completed and ░ remaining.
  • Round down to the nearest 5% unless completion evidence supports 100%.
  • Show only the overall bar unless the user requests subtask bars.
  • In multi-agent work, only the coordinator shows the overall bar.
  • The progress line is the final content in the update. End immediately after it with no note, recap, or closer.

Persist a preference

The greenline CLI installs this skill; there is no global installation step here and no machine-global instruction write. A request to change how or how often you report is followed within the current task and saved nowhere. Persist a preference only when the owner asks for a lasting repository rule, as a house ruling in the repository's user-owned instruction home, in the format the work guide defines. The next context reads that rule without demanding a second confirmation.

Durable output: none. Applied on its own cadence while long work runs, and the moment the owner asks to change how or how often you report, before any preference is saved anywhere.

04 同仓库更多 Skills

fern-works
fern-works/greenline

use-garden

Consulting garden, the separately installed guidance command this repository's owner enabled, when an open engineering choice the repository does not settle, made in a change or only recommended, would be informed by published guidance. Use for a choice between approaches, a pattern or pitfall for a technology or concern the work touches, or a unit a repository decision names; not for a question about this repository's own facts or a change whose choices are already settled.

★ 0
fern-works
fern-works/greenline

use-garden

Consulting garden, the separately installed guidance command this repository's owner enabled, when an open engineering choice the repository does not settle, made in a change or only recommended, would be informed by published guidance. Use for a choice between approaches, a pattern or pitfall for a technology or concern the work touches, or a unit a repository decision names; not for a question about this repository's own facts or a change whose choices are already settled.

★ 0
fern-works
fern-works/greenline

what-to-build

Propose what this product could do next at three altitudes: sharpen a shipped slice, extend one, or open ground the product does not stand on yet. Use for 'what should we build next', 'any ideas for this app', 'what's missing', 'is this done or is there more in it', and quick-win hunts. Not for refactors, which belong to the architecture survey.

★ 0
fern-works
fern-works/greenline

wayfinder

Plan a huge chunk of work (more than one agent session can hold) as a shared map of decision tickets in a plan directory's map.md, and resolve them one at a time until the way to the destination is clear.

★ 0
fern-works
fern-works/greenline

verify

Prove a change's claims with fresh evidence: restate each item of its acceptance so it can fail, capture a baseline and a treatment with the same command, compare the raw artifacts, and return VERIFIED, NOT VERIFIED or INCONCLUSIVE for each, at the depth the change's mode gives the verify step. Use when the mode reaches verify, or when the owner says "verify this", "prove it works" or "did this fix it".

★ 0
fern-works
fern-works/greenline

tdd

Test-first development at any step: discover the repository's own test commands, name the break a test catches, watch it fail for the intended reason, make it pass with the smallest change, and add a new test only where it catches a realistic failure the existing ones would miss. Use when writing or changing a test, building a feature or fixing a bug test-first, or deciding whether a change needs a new test; also on 'red-green-refactor' and integration tests.

★ 0