ExecutePlan
Execute an implementation plan inline — task by task in a single session — then finish the work: verify, present integration options (merge, PR, keep…
Skill by N4M3Z
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/n4m3z-executeplan
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.
About
Use cases: productivity
Reviews
Write a reviewNo 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/executeplan
What's inside
- Skill name
ExecutePlan- Description
- Execute an implementation plan inline — task by task in a single session — then finish the work: verify, present integration options (merge, PR, keep, discard), and clean up the workspace. USE WHEN executing plan, run plan, implement plan, inline execution, implementation complete, deciding merge strategy, finishing a feature branch or workspace.
- License
- EUPL-1.2
- Instructions
- 66 lines in SKILL.md · view on GitHub
Agent Skills spec issues
- name: only lowercase letters, digits and single hyphens allowed
Files (1)
- SKILL.md
SKILL.md (first lines)
# ExecutePlan
Execute a Plan task-list output in the current session, one task at a time, then land the result. This is the sequential alternative to DeveloperSprint (parallel agents). Use ExecutePlan for smaller plans or when tasks have sequential dependencies.
1. **Load the plan** — read the plan file from `docs/plans/`
2. **Critical review** — before starting, check: are tasks ordered correctly? Are dependencies satisfied? Any gaps? If issues found, fix the plan first or ask the user.
3. **Set up workspace** — if the plan modifies existing code, isolate by VCS: a jj workspace (`jj workspace add ../repo-<name>`) in jj-colocated repos, a git worktree in git-only repos. Run project setup and the test suite once for a clean baseline; stop on a broken baseline.
4. **Execute task by task**:
- Read the task specification
- Implement exactly what the task says
- Run the verification step specified in the task
- Invoke VerifyCompletion before marking the task done
- Move to the next task only after the current one passes
5. **Handle blockers** — if stuck, stop and ask the user. Never guess past a blocker.
6. **Final verification** — after all tasks, run the full build + test suite. If tests fail, fix before offering integration; never merge or PR a red tree.
7. **Finish** — present exactly these options and execute the choice:
1. Merge back to the base branch locally
2. Push and create a Pull Request
3. Keep as-is (user handles it later)
4. Discard this work
8. **Report** — summarize what was done, what was skipped, what needs follow-up
## Finishing mechanics
Git-only repos:
- **Merge locally:** `git switch <base>` → `git pull` → `git merge <feature>` → re-run tests on the result → `git branch -d <feature>`
- **PR:** `git push -u origin <feature>` → `gh pr create` (or `glab`) with Summary + Test plan body
- **Discard:** requires the user to type `discard` — it deletes the branch, its commits, and the worktree. Then `git switch <base>` → `git branch -D <feature>`
- **Cleanup (merge and discard; PR keeps the tree until merged):** leave the worktree (`cd` to the main checkout), `git worktree remove <path>`, then sweep orphans: any linked worktree in `git worktree list` whose tree is clean and whose HEAD is already an ancestor of mainline gets removed. Session-scoped worktrees sit at detached HEAD under `~/worktrees/<repo>/<name>/`, invisible to branch-based checks.
jj-colocated repos (no branches to finish, no worktrees to remove):
- **Land:** `jj bookmark set <name> -r @-` on the described change, then `jj push` (the gated alias)
- **Discard:** `jj abandon` the change (op-log-reversible, no typed confirmation needed)
- **Cleanup:** `jj workspace forget <name>` for any workspaces created in step 3; the directory is removed separately
## Red Flags
| Thought | Reality |
| ------------------------------------------------- | ------------------------------------------------------------ |
| "I'll do tasks 3 and 4 together, they're related" | One task at a time. Verify each before moving on. |
| "This task is wrong, let me improvise" | Stop. Fix the plan or ask the user. Don't freelance. |
| "I can skip verification on this one" | No. VerifyCompletion on every task. |
| "I'm blocked but I can probably work around it" | Stop and ask. Workarounds create hidden dependencies. |
| "Let me also fix this while I'm here" | Scope creep. Do what the plan says, nothing more. |
| "The plan is mostly done, close enough" | Partially executed plans are worse than unstarted ones. |
| "Tests are red but the merge is small" | Never merge or PR a red tree. Fix first. |
| "I'll clean up the worktree later" | A merge is not complete while its worktree still exists. |
## Constraints
- One task at a time — never batch or skip ahead
- Verify each task before proceeding to the next
- Stop on blockers — ask the user, don't guess
- No scope expansion beyond the plan
- Isolate by VCS: jj workspace in jj-colocated repos, git worktree in git-only repos
- Present exactly the four finish options; typed `discard` confirmation before destroying git work
- For plans with independent tracks, use DeveloperSprint instead
## Sources
- <https://github.com/obra/superpowers>
- <https://github.com/davila7/claude-code-templates>From N4M3Z/forge-core, EUPL-1.2.
More from N4M3Z/forge-core
- 0.520score
LocalModel
N4M3Z
Run prompts against a local LLM backend (LM Studio, Ollama, or MLX) with privacy-preserving output guardrails — drop reasoning_content, enforce structured…
SkillCFreshness: C★ 4Claude Code +9 - 0.520score
ForgeSkill
N4M3Z
Create, validate, evaluate, and iterate skills for forge modules.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
Brainstorming
N4M3Z
Collaborative ideation before implementation — explore context, clarify intent, propose approaches.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildAgent
N4M3Z
Create, validate, or audit agent definitions.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildHook
N4M3Z
Create and validate module hooks.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildModule
N4M3Z
Design, build, and validate forge modules.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildPlugin
N4M3Z
Create, validate, and publish Claude Code plugins from forge modules.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildRule
N4M3Z
Create, validate, or place behavioral rules in forge modules.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildSkill
N4M3Z
Create and validate skills for forge modules.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
DesignFrontend
N4M3Z
Build distinctive, production-grade frontend interfaces that avoid generic AI aesthetics.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
DesignSpec
N4M3Z
Write a formal design spec from brainstorming output or requirements.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
GuardRails
N4M3Z
Safety guardrails: runtime overrides for AI-safety plugin blocks, and per-language security review for python, javascript/typescript, and go.
SkillCFreshness: C★ 4Claude Code +9
Signals
- Stars
- 4
- Forks
- 3
- Open issues
- 8
- Last commit (this folder)
- Jul 13, 2026 (3 months ago)
- Last release
- Apr 16, 2026 (6 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- EUPL-1.2
- Repository
- github.com/N4M3Z/forge-core
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from N4M3Z
- 0.520score
LocalModel
N4M3Z
Run prompts against a local LLM backend (LM Studio, Ollama, or MLX) with privacy-preserving output guardrails — drop reasoning_content, enforce structured…
SkillCFreshness: C★ 4Claude Code +9 - 0.520score
ForgeSkill
N4M3Z
Create, validate, evaluate, and iterate skills for forge modules.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
HtmlPlayground
N4M3Z
Generate single-file HTML demos comparing techniques side-by-side.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildAgent
N4M3Z
Create, validate, or audit agent definitions.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
BuildHook
N4M3Z
Create and validate module hooks.
SkillCFreshness: C★ 4Claude Code +9 - 0.515score
Brainstorming
N4M3Z
Collaborative ideation before implementation — explore context, clarify intent, propose approaches.
SkillCFreshness: C★ 4Claude Code +9
Cite this page
ExecutePlan skill. Agent Mart. https://agentmart.dev/skills/n4m3z-executeplan (last updated 2026-10-24)