How do I install the Scope And Execute skill?

Run npx skillmds@latest add forge-space/scope-and-execute in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Scope And Execute safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Scope And Execute?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Scope And Execute free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Scope And Execute?

Forge-Space (@forge-space) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Scope And Execute by forge-space — Agent Skill

Name: Scope And Execute
Author: Forge-Space

Scope And Execute

Composite skill — understand a problem area, plan the work, execute it, and ship it in one chained workflow. Chains adt-repo-intake or ecosystem-health (analysis) → context-pack (load context) → plan (phased plan) → dispatch or loop (execute) → ship (merge). Use when the task is "fix the X area" or "implement Y end-to-end" — the single command for analysis through merge.

Forge-Space 6b26448 2 files · 4.2 KB Updated Sep 17, 2026

File contents

Scope and Execute

Replaces the "should I /plan first or /loop?" decision with one workflow that handles the whole arc: understand → plan → execute → merge.

Auto-invocation triggers

User asks to "fix X", "implement Y", "build the Z feature" with non-trivial scope
User asks to "look at this repo and figure out what to do"
Work spans multiple files or services
Estimated >1h of effort

Workflow

Phase 1 — Analysis (always)

New repo / unfamiliar code: invoke adt-repo-intake
Multi-repo or workspace context: invoke ecosystem-health
Already in active work: skip Phase 1
Output: scope brief — which files, which contracts, which dependencies

Phase 2 — Context loading (always for non-trivial work)

Invoke context-pack (or auto-context-pack if it has already fired this prompt)
Pulls relevant code, standards, prior ADRs, related decisions
Cap at 2k tokens

Phase 3 — Plan (skip for trivial single-file changes)

Invoke plan to write a phased plan with validation per phase
Output: .claude/plans/<task>.md
Skip if scope is genuinely 1-2 files with obvious edits

Phase 4 — Execute (always)

For independent parallel tracks: invoke dispatch
For sequential inspect→act→verify: invoke loop
For complex 3+ phase work: invoke three-man-team
Implement against Phase 3's plan

Phase 5 — Pre-merge (always for shippable work)

Invoke merge-confidently (or pr-merge-readiness directly)
Resolve any blockers
Invoke ship when verdict is MERGE

Phase 6 — Capture (always)

If non-trivial decision was made: invoke adr-write
Invoke knowledge-loop to update memory + RAG

Reconciliation

Output a scope summary at start, status updates between phases, final summary:

SCOPE AND EXECUTE — <task>
  Phase 1 Analysis:    <files in scope>, <key dependencies>
  Phase 2 Context:     <token count loaded>
  Phase 3 Plan:        <plan path, N phases>
  Phase 4 Execute:     <commits, files changed>
  Phase 5 Ship:        <PR number, merge SHA>
  Phase 6 Capture:     <ADR / memory paths>

Outputs / Evidence

Scope brief
Plan file
Implementation commits
Merged PR
Decision capture

Failure / Stop Conditions

Analysis reveals scope >1 week of work → stop, recommend breaking into phases
Plan rejected by user → restart Phase 3 with their feedback
Phase 5 verdict is FIX → stop and surface blocker, do not push through
Never auto-merge without Phase 5 returning MERGE

---name: scope-and-executedescription: Composite skill — understand a problem area, plan the work, execute it, and ship it in one chained workflow. Chains adt-repo-intake or ecosystem-health (analysis) → context-pack (load context) → plan (phased plan) → dispatch or loop (execute) → ship (merge). Use when the task is "fix the X area" or "implement Y end-to-end" — the single command for analysis through merge.---# Scope and ExecuteReplaces the "should I /plan first or /loop?" decision with one workflow that handlesthe whole arc: understand → plan → execute → merge.## Auto-invocation triggers- User asks to "fix X", "implement Y", "build the Z feature" with non-trivial scope- User asks to "look at this repo and figure out what to do"- Work spans multiple files or services- Estimated >1h of effort## Workflow### Phase 1 — Analysis (always)- New repo / unfamiliar code: invoke `adt-repo-intake`- Multi-repo or workspace context: invoke `ecosystem-health`- Already in active work: skip Phase 1- Output: scope brief — which files, which contracts, which dependencies### Phase 2 — Context loading (always for non-trivial work)- Invoke `context-pack` (or `auto-context-pack` if it has already fired this prompt)- Pulls relevant code, standards, prior ADRs, related decisions- Cap at 2k tokens### Phase 3 — Plan (skip for trivial single-file changes)- Invoke `plan` to write a phased plan with validation per phase- Output: `.claude/plans/<task>.md`- Skip if scope is genuinely 1-2 files with obvious edits### Phase 4 — Execute (always)- For independent parallel tracks: invoke `dispatch`- For sequential inspect→act→verify: invoke `loop`- For complex 3+ phase work: invoke `three-man-team`- Implement against Phase 3's plan### Phase 5 — Pre-merge (always for shippable work)- Invoke `merge-confidently` (or `pr-merge-readiness` directly)- Resolve any blockers- Invoke `ship` when verdict is MERGE### Phase 6 — Capture (always)- If non-trivial decision was made: invoke `adr-write`- Invoke `knowledge-loop` to update memory + RAG## ReconciliationOutput a scope summary at start, status updates between phases, final summary:```SCOPE AND EXECUTE — <task> Phase 1 Analysis: <files in scope>, <key dependencies> Phase 2 Context: <token count loaded> Phase 3 Plan: <plan path, N phases> Phase 4 Execute: <commits, files changed> Phase 5 Ship: <PR number, merge SHA> Phase 6 Capture: <ADR / memory paths>```## Outputs / Evidence- Scope brief- Plan file- Implementation commits- Merged PR- Decision capture## Failure / Stop Conditions- Analysis reveals scope >1 week of work → stop, recommend breaking into phases- Plan rejected by user → restart Phase 3 with their feedback- Phase 5 verdict is FIX → stop and surface blocker, do not push through- Never auto-merge without Phase 5 returning MERGE

Frequently asked questions