agentleFS
skyf0xxskyf0xx/gambit/skills/brief/SKILL.md

brief

Skill29 starsChanged 33 days agoMITView on GitHub

Use when the user wants to understand where things stand in plain language rather than in framework terms — returning after a gap, feeling lost, showing the situation to someone else, or asking "so what does all this actually mean". Translates the goal into ordinary prose, explains any jargon it contains, and names the one thing to do next. Also gives a terse quick snapshot for "where are we" or "are we on track". Read-only.

We split this skill into the rules it gives an agent

3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 68 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read-only. Never writes to the goal.

    line 23
  2. Distinctive:

    Use instead of the full brief when the ask is short: "where are we", "are we on track", "quick status".

    line 29
  3. Distinctive:

    Not dumbed down — a competent adult who happens not to know strategic planning vocabulary.

    line 56

“Common” is measured: kept by at least 6 in 10 of the 68 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill: brief

Trigger: The user wants to know where things stand, in plain terms. Returning after a break, feeling lost in the analysis, needing to explain the situation to somebody else, or asking some version of "so what does all this actually mean".

Also use when the goal has accumulated framework vocabulary — Schwerpunkt, CoG, CV, culminating point — and the user shouldn't have to hold a glossary to read their own goal.

Purpose: Translate. Give the picture the way you'd explain it to a smart friend who has no background in any of it.

Read-only. Never writes to the goal.


Quick Snapshot Mode

Use instead of the full brief when the ask is short: "where are we", "are we on track", "quick status". Terse, numbers and states, no interpretation. Read-only, like the rest of this skill.

SNAPSHOT [date]
GOAL      [the goal sentence] · deadline [date or none] · [N days left]
POSTURE   [level and label — omit if none]
FOCUS     [the current focus, or "none set"]
PLAN      [one line per line of operation: label · status · next move and its date]
OVERDUE   [moves past their date — omit if none]
OPEN      [open questions, forks and unscored forecasts, counted — omit if none]
LAST EVAL [the date and the verdict, or "never run"]
RECENT    [the last 3 log entries, one line each]

For "are we on track", answer from what the record shows: the latest assessment, the criteriaStatus counts, the overdue list. If the record is thin or mixed, say it is unclear and point to eval for a real audit. Don't soften a stall to be polite.


Voice & Tone

Plain, direct, unhurried. Short sentences. No framework terms unless you immediately say what they mean in ordinary words.

Not dumbed down — a competent adult who happens not to know strategic planning vocabulary. Respect the intelligence, drop the jargon.

Honest about trouble. If things are stalled, say stalled — translating into accessible language must never soften a hard finding. That's the one way this skill does real damage.


Execution Sequence

1. Load Context

The goal's current state is already supplied in "Current goal state." Call get_goal instead if the user may have edited the dashboard since.

2. Write the Brief

Prose, not a template. Roughly this shape, in this order, adapted to what's actually in the goal:

What you're trying to do — the goal, in one or two sentences, as a person would say it out loud.

How you'll know it worked — the success criteria, restated concretely. If a criterion isn't something the user can directly cause — a government decision, another organisation's choice — say so here. That's not a flaw, but it changes what progress looks like, and it should be visible.

Where things actually stand — what's moved, what hasn't. Pull from the log. Be specific about dates. If nothing has moved in a while, say that.

What you're focused on right now, and why — the current Schwerpunkt in plain terms. Say the word once, in parentheses, if it's on the dashboard — the user will see it there and should know what it means.

What could go wrong — the top one or two risks from the riskNotes key or the systemsNotes key, translated. Not the whole list. The ones that would actually hurt.

Who's involved — if the people key is non-empty: who's committed, who's not confirmed yet, and who the effort would be in trouble without.

The clock — deadline, time remaining, and a straight answer on whether the current pace fits. Yes, no, or genuinely unclear.

3. Translate the Jargon

Whenever a framework term appears in the goal, define it inline the first time, in one clause:

  • Schwerpunkt — the one thing worth concentrating on right now
  • Centre of Gravity (CoG) — what a side's strength actually depends on
  • Critical Vulnerability (CV) — the weak point that, if it gives, takes the strength with it
  • Culminating point — the moment past which more effort starts producing less
  • Line of effort (LOE) — one ongoing workstream running toward the goal
  • Posture — how hard you're currently pushing, and how much risk that carries
  • PMESII / ASCOPE — checklists for reading an environment without missing a dimension
  • Admiralty rating (e.g. B/2) — how trustworthy a source is, and how well-corroborated the specific claim is

Define only what's actually present. Don't teach the glossary.

4. Name the Next Thing

End with one concrete next action and a short menu. Not a list of everything possible — the single most useful move, then the alternatives:

The most useful thing you could do next: [one specific action]

If you'd rather:
  - Rethink what to focus on → strategy
  - Turn this into steps → plan
  - Find out something you're unsure about → web search
  - Make a call you've been putting off → decide
  - Check whether this is really working → eval

5. Offer to Go Deeper

Close by inviting the question the brief probably raised:

Anything in there you want the longer version of?

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.