Skip to content
Agentmart
Submit a tool

rework

Rework alignment and lightweight definition for an intentional implementation change that must not degrade measured outcomes.

Skill by marcuskrogh

IndexedMIT★ 1
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/marcuskrogh-rework

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Rework alignment and lightweight definition for an intentional implementation change that must not degrade measured outcomes. Produces REWORK.md and one tracker Task with a parity-iterative binding. Prefer /define for new work (agent classifies as rework).

Use cases: productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/rework

What's inside

Skill name
rework
Description
Rework alignment and lightweight definition for an intentional implementation change that must not degrade measured outcomes. Produces REWORK.md and one tracker Task with a parity-iterative binding. Prefer /define for new work (agent classifies as rework).
License
MIT
Instructions
93 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Rework

Applies [CONCEPT_ALIGNMENT](../concepts/CONCEPT_ALIGNMENT.md) and
[CONCEPT_DEFINITION](../concepts/CONCEPT_DEFINITION.md) to a **bounded rework**.
Produces `REWORK.md`. Comparative verification notes:
[../implement/rework.md](../implement/rework.md). Shared persist/track contract:
[../define/overrides.md](../define/overrides.md). Applies
[CONCEPT_SKILL](../concepts/CONCEPT_SKILL.md).

**On invoke:** read [../define/overrides.md](../define/overrides.md) and
[../implement/rework.md](../implement/rework.md).

## Extensions

| Slot | This skill |
|------|------------|
| **Subject** | Limited area whose implementation should change (algorithm, control law, mapping, internal path) while measured outcomes stay within a declared bar |
| **Probes** | Area boundary; thin description of current vs intended implementation; why rework; **parity bar** (metrics, scenarios, tolerances, how baseline is obtained); **pass criteria**; out of scope; optional parent Story/Task link; optional `/model` when math equivalence needs alignment |
| **Stop condition** | The user has stated the area, intended rework, parity bar, and pass criteria (or explicitly deferred one) and approved the readiness prompt |
| **Alignment / definition artifact** | `REWORK.md` (path from WORKSPACE) |
| **Readiness prompt** | "Is this enough to implement the rework?" |
| **Opening** | Thin description **required**. Missing area → "What area should be reworked?" Rich (area + change pasted): first question on highest-impact gap (often the parity bar) |
| **Scope guard** | Bounded area; intentional implementation change allowed; **no silent degradation** — parity bar required; keep the define loop short (richness lives in implement comparative eval) |
| **Depth** | Lightweight |
| **Work packages** | Optional Sub-tasks only when packages are truly separate |

## Artifact

```markdown
# Rework: [title]

## Summary
- …

## Area
- …   # module, controller, mapping, internal path, …

## Current implementation
- …

## Intended implementation
- …   # e.g. replace PID with static map

## Why rework
- …

## Parity bar
- Metrics: …
- Scenarios / fixtures: …
- Tolerances / “worse” definition: …
- Baseline: …   # how current code is measured (branch, seam, recorded runs)
- Candidate: …  # how new code is measured on the same scenarios

## Pass criteria
- …   # include: candidate meets parity bar vs baseline; suite green

## Out of scope
- …   # unrelated behaviour changes, unrelated areas, …

## Work packages
1. …   # optional; omit if a single package is enough

## Tracker
- Task: <KEY>
- Sub-tasks: … (if any)
- Branch: <delivery-branch>
- PR: <url or draft url>

## Classification
- Class: rework
- Confidence: high
- Why: explicit /rework

## Workflow
- Template: parity-iterative
- Parameters: catalog defaults for parity-iterative (`implement.verify=comparative`, `implement.iteration=until-bar`)
- Chain: delivery chain for that template
```

## Inputs

| Input | When present | When absent |
|-------|----------------|-------------|
| Area and intended implementation change | Apply them | Ask "What area should be reworked?" once |
| Parity bar | Apply it | Align the bar in this invocation before the artifact is approved |

## Output

`REWORK.md` — rework spec, parity bar, and the parity-iterative binding. Outcome: `ready`.

This skill does not name a successor. Apply the workflow transition before the turn ends.

From marcuskrogh/skills, MIT.

More from marcuskrogh/skills

Signals

Stars
1
Forks
1
Open issues
0
Last commit (this folder)
Sep 27, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from marcuskrogh

Cite this page

rework skill. Agent Mart. https://agentmart.dev/skills/marcuskrogh-rework (last updated 2026-10-24)