mcpField
E
explain-outcome

explain-outcome

github.com/lwlee2608Agent workflows180 installsUpdated 1d ago
Repository

About explain-outcome

Explain Outcome is a writing skill for reporting results in three ordered sections: what happened, what it means, and what's next. It separates facts from marked guesses and states a concrete action or that none is needed.

Key features

  1. 01Three-part outcome format
  2. 02Separates facts from guesses
  3. 03Marks unconfirmed interpretations
  4. 04States the next action
  5. 05Allows a one-sentence small-result reply

Use cases

  1. 01Summarizing test or command results
  2. 02Explaining an outcome without mixing facts and guesses
  3. 03Reporting uncertainty clearly
  4. 04Closing a small task with no decision

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/lwlee2608/agent-skills --skill explain-outcome

After installing, type/explain-outcomein Claude Code to confirm it loaded.

Details

From the project README

Explain Outcome

Report the result in three parts, in this order. The goal is only this: a human reads it once and understands.

Write in ASD-STE100 Simplified Technical English. Short sentences are easier to read than long ones. The active voice is easier than the passive. One word for one meaning is easier than a synonym each time. Plain words are easier than jargon.

Format

**What happened**
<the facts: what ran, what it returned, what changed>

**What it means**
<the cause, the effect, and what is not known yet>

**What's next**
<the action the agent will do, or the decision the user must make>

Facts go under "What happened". Guesses go under "What it means", marked "likely" or "not confirmed". "What's next" is an action: "I will..." or "You must decide...". If no action is necessary, write "Nothing. Done." For a small result with no decision, reply in one sentence instead.

Worked example

User: "Run the tests."

Good:

**What happened**
`make test` failed. 1 of 42 tests failed:

    FAIL: TestOrderTotal_Discount (order_test.go:88)
    expected 90.00, got 100.00

**What it means**
The discount is not in the order total. The likely cause is an early return in `applyDiscount` (order.go:41). This is not confirmed. No other test failed.

**What's next**
- I will read `applyDiscount` to confirm the cause, then fix it.
- You must decide: does a 0% discount call `applyDiscount`, or skip it?

Bad — facts and guesses in one sentence, asks instead of states:

Tests failed because the discount logic is broken in order.go, probably the early return. I can fix it if you want.

Comments

Markdown supported