Execute Implementation Plans with Checkpoints
Implementation plans can lose focus when work moves into a separate AI session. This skill guides Claude, Codex, and Claude Code through batch execution, verification, and review checkpoints.
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Review the Skillstore skill "executing-plans" from https://skillstore.io/skills/obra-executing-plans.md and its manifest at https://skillstore.io/api/skills/obra-executing-plans/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.Your Agent should still show its plan and request any confirmation required by the security policy.
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Using "executing-plans". Run the first batch from the migration plan.
Expected outcome:
Using "executing-plans". Continue after my review comments.
Expected outcome:
I reviewed SKILL.md, the only scanned file, and found prompt-only workflow instructions. Static analysis reported no suspicious patterns, and I found no semantic evidence of prompt injection, data exfiltration, or business-logic abuse.
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
https://skillstore.io/skills/obra-executing-plans/audits/8?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report[](https://skillstore.io/skills/obra-executing-plans?utm_source=security_passport_badge)<a href="https://skillstore.io/skills/obra-executing-plans?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/obra-executing-plans/security.svg" alt="Skillstore security assessment" loading="lazy"></a><iframe src="https://skillstore.io/embed/skills/obra-executing-plans.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>obra. (2026). executing-plans security audit report (audit version 8) [Author version unspecified]. Skillstore. https://skillstore.io/skills/obra-executing-plans/audits/8@techreport{obra-obra-executing-plans-2026,
author = {obra},
title = {executing-plans security audit report (audit version 8)},
institution = {Skillstore},
year = {2026},
number = {8},
url = {https://skillstore.io/skills/obra-executing-plans/audits/8},
note = {Author version unspecified}
}cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "executing-plans security audit report (audit version 8)"
version: "unspecified"
type: report
authors:
- name: "obra"
date-released: "2026-07-06"
url: "https://skillstore.io/skills/obra-executing-plans/audits/8"
identifiers:
- type: other
value: "skillstore:obra-executing-plans:audit:8"
description: "Skillstore immutable audit report identifier"
Each author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.
Why this variant is first
obra-executing-plans
2026-08-21
davila7-executing-plans
2026-08-21
codingcossack-executing-plans
2026-08-21
sickn33-executing-plans
2026-08-21
cygnusfear-executing-plans
2026-08-21
doyajin174-executing-plans
2026-08-21
cycleaddict-executing-plans
2026-08-21
dmjgilbert-executing-plans
2026-08-21
zhanlincui-executing-plans
2026-08-21
dyai2025-executing-plans
2026-08-21
Use it to help an assistant execute an approved technical plan in small, reviewable batches.
Use it to keep implementation focused when returning to a plan in a new session.
Use it when a reviewer wants progress reports and verification output before more tasks proceed.
Use the executing-plans workflow on docs/implementation-plan.md. Review the plan first, then complete the first batch.
Follow the plan in PLAN.md. Stop before work if any task lacks verification, dependencies, or ownership.
Continue executing the next batch from the existing plan. Apply my feedback first, then report implemented changes and verification results.
Execute the remaining plan batches with review checkpoints. When all tasks pass verification, transition to the finishing branch workflow.
Author
obraLicense
MIT
Skillstore revision
r1
Version notice
The author did not declare a version.
Ref
e397ce5a135369cd6def89cb6400d013e2f68f2f
Maintenance freshness
7/18/2026
Usage
8 downloads · 215 views
File structure
📄 SKILL.md
Create Detailed Implementation Plans
Complex work is risky when requirements are not broken into clear steps. This skill turns specs into test-focused implementation plans for Claude, Codex, and Claude Code.
Build Tested AI Skills
AI skills can fail when instructions are vague or untested. This skill guides test-driven skill writing with pressure scenarios and verification loops.
Evaluate Code Review Feedback
Code review feedback can be unclear, incorrect, or incomplete. This skill helps Claude, Codex, and Claude Code verify suggestions, ask precise questions, and implement reviewed changes carefully.
Plan Features Through Guided Brainstorming
Feature ideas often move to implementation before goals and constraints are clear. This skill guides Claude, Codex, and Claude Code through questions, trade-offs, and validated design notes.
Execute Plans with Subagent Reviews
Complex implementation plans lose quality when one context handles every task. This skill coordinates fresh subagents for implementation, spec checks, and code quality reviews.
Request Code Reviews Before Merge
Unreviewed changes can ship bugs, regressions, or incomplete requirements. This skill guides Claude, Codex, and Claude Code to request structured review before merge.
Create Verifiable Agent Loop Specs
by zhangzeyu99-web
Vague coding-agent tasks drift, repeat work, and report success without evidence. Ariadne Loop turns them into bounded loop specs with verifiers, stop rules, and human gates.
Create Focused Implementation Plans
by sickn33
Multi-step coding work can become vague, large, and hard to verify. This skill helps Claude, Codex, and Claude Code create short plans with clear tasks, order, and checks.
Verify Completion Claims Before Reporting
by CodingCossack
Completion claims can be wrong when agents rely on memory or confidence. This skill requires fresh evidence before reporting success.
Implement Conductor Tracks Reliably
by sickn33
Conductor tracks can drift when tasks, tests, and metadata are updated by hand. This skill guides Claude, Codex, or Claude Code through TDD implementation, checkpoints, commits, and progress updates.
Verify Agent Completion With Evidence
by CleanExpo
AI agents can mark work complete without proof, which creates false completion risk. This skill requires independent evidence before task completion is accepted.
Run Consistent Git Workflows
by Doyajin174
Git work can drift into unclear commits and missed verification. This skill guides conventional commits, branch naming, checks, and push preparation.