agentsclimarketplace

Explain

Skill hereinthehive/gotrino-assistant/skills/explain

Document the reasoning behind implementation decisions as markdown files. Creates decision records that future developers can reference.From its SKILL.md

Install
npx -y skills add hereinthehive/gotrino-assistant --skill explain

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.0 KB, as published. Nobody here has run it

Explain (Decision Documentation)

Create decision records that document the reasoning behind implementation choices.

Philosophy

Good code is self-documenting for what it does. But why it does it that way? That's lost to time unless we capture it.

"Six months from now, when someone asks 'Why did we do it this way?', will there be an answer?"

The value of this skill isn't compliance—it's preserving context that makes the codebase maintainable.

Config Integration

Follow the migration preflight in references/config-migration.md, then check for .assistant-config.md in the project root.

Decisions location (in order of precedence):

  1. Decisions location setting in config
  2. Existing decisions/ directory
  3. Existing docs/decisions/ or docs/adr/ directory
  4. Default: create decisions/

If no config exists, ask the user where to store decisions or use the default.

When to Use

Run /explain after:

  • Completing a significant implementation
  • Making an architectural decision
  • Choosing between multiple valid approaches
  • Implementing something counterintuitive
  • When future developers might ask "why was it done this way?"

Input

The user will provide:

  • File path(s) to explain, OR
  • Feature/change description, OR
  • Specific question ("Why did we use X instead of Y?")

If nothing provided, explain recent uncommitted changes.

Process

1. Understand the Decisions

For the code being explained, identify:

  • Key architectural choices
  • Algorithm or pattern selections
  • Library/dependency choices
  • Trade-offs that were made
  • Constraints that influenced the approach

2. Research Context

Gather context before writing:

  • What patterns exist elsewhere in the codebase?
  • Were there constraints from existing code?
  • What does the task file or PR description say?
  • What alternatives could have been used?

3. Generate File Name

Use the pattern: NNNN-short-description.md

  • Check existing files in decisions directory
  • Use next available number (0001, 0002, etc.)
  • Keep description short (2-4 words, kebab-case)

Examples:

  • 0001-use-zustand-for-state.md
  • 0002-jwt-over-sessions.md
  • 0003-monorepo-structure.md

4. Write Decision Record

Create the file in the decisions directory.

Output Format

# [Number]. [Decision Title]

**Date**: [YYYY-MM-DD]
**Status**: Accepted
**Context**: [What prompted this decision]

## Decision

[1-2 sentence summary of what was decided]

## Why This Approach

- [Primary reason]
- [Secondary reason]
- [Constraint that influenced this]

## Alternatives Considered

### [Alternative A]

**Pros**: [benefits]
**Cons**: [drawbacks]
**Why not**: [specific reason rejected]

### [Alternative B]

**Pros**: [benefits]
**Cons**: [drawbacks]
**Why not**: [specific reason rejected]

## Trade-offs Accepted

- [What we gave up and why it's acceptable]

## Consequences

- [What this decision means for future development]
- [Things to watch out for]

## Related

- [Links to related code, PRs, or other decisions]

After Writing

Output confirmation:

## Decision Record Created

**File**: `decisions/0003-use-zustand-for-state.md`

### Summary
Documented decision to use Zustand over Redux for state management.

### Key Points
- Simpler API reduces boilerplate
- Trade-off: Less ecosystem tooling
- Revisit if: Complex state debugging needed

Why This Matters

Decision records serve multiple purposes:

  • Onboarding: New developers understand architectural choices
  • Maintenance: Future changes can respect original constraints
  • Accountability: Decisions are traceable and reviewable
  • Learning: The team builds shared understanding over time

Without this context, teams make the same decisions repeatedly, or worse—undo good decisions because they don't understand why they were made.

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.