Refresh AI Focus During Flow Tasks
Long coding sessions can push original goals out of the model attention. This skill adds refresh checkpoints before key workflow decisions.
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Review the Skillstore skill "flow-attention-refresh" from https://skillstore.io/skills/dimon94-flow-attention-refresh.md and its manifest at https://skillstore.io/api/skills/dimon94-flow-attention-refresh/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 "flow-attention-refresh". Start the next task after a long conversation.
Expected outcome:
The assistant rereads the current task, restates the goal, names the acceptance criteria, and begins only after confirming alignment.
Using "flow-attention-refresh". A build failed during a TDD loop.
Expected outcome:
The assistant records the failure, reviews related error history, rereads the task definition, and selects a focused recovery step.
Using "flow-attention-refresh". Continue an autonomous Ralph loop.
Expected outcome:
The assistant finds the next unfinished task, checks recent errors, states the immediate action, and avoids repeated mistakes.
All external command findings are false positives from Markdown code fences and inline backticks in SKILL.md. The high-entropy finding is also a false positive; the file is readable documentation with no encoded payload or prompt injection evidence.
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/dimon94-flow-attention-refresh/audits/8?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report[](https://skillstore.io/skills/dimon94-flow-attention-refresh?utm_source=security_passport_badge)<a href="https://skillstore.io/skills/dimon94-flow-attention-refresh?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/dimon94-flow-attention-refresh/security.svg" alt="Skillstore security assessment" loading="lazy"></a><iframe src="https://skillstore.io/embed/skills/dimon94-flow-attention-refresh.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Dimon94. (2026). flow-attention-refresh security audit report (audit version 8) [Author version unspecified]. Skillstore. https://skillstore.io/skills/dimon94-flow-attention-refresh/audits/8@techreport{dimon94-dimon94-flow-attention-refresh-2026,
author = {Dimon94},
title = {flow-attention-refresh security audit report (audit version 8)},
institution = {Skillstore},
year = {2026},
number = {8},
url = {https://skillstore.io/skills/dimon94-flow-attention-refresh/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: "flow-attention-refresh security audit report (audit version 8)"
version: "unspecified"
type: report
authors:
- name: "Dimon94"
date-released: "2026-07-06"
url: "https://skillstore.io/skills/dimon94-flow-attention-refresh/audits/8"
identifiers:
- type: other
value: "skillstore:dimon94-flow-attention-refresh:audit:8"
description: "Skillstore immutable audit report identifier"
Use refresh checkpoints before each task so the agent restates the current goal and acceptance criteria.
Apply the same document reread step whenever a flow command starts or changes direction.
Reread recent error notes and the active task before choosing a fix after a failure.
Use flow-attention-refresh before this task. Read the current TASKS.md task and DoD, then summarize the acceptance criteria.
A test or build failed. Use the error recovery protocol, review ERROR_LOG.md and the active task, then propose the next fix.
Start the next Ralph iteration. Refresh from the first unfinished TASKS.md item and recent ERROR_LOG.md entries before acting.
Apply this skill to a new flow command. Define when to reread goal files and what each checkpoint must focus on.
Author
Dimon94License
MIT
Skillstore revision
r1
Version notice
The author did not declare a version.
Ref
72d5025b022c77f7a51bdf5c1637c689c80e89d1
Maintenance freshness
7/18/2026
Usage
7 downloads · 211 views
File structure
📄 SKILL.md
Generate Directory CLAUDE.md Guides
Directory changes often leave local documentation outdated. This skill gives Claude, Codex, and Claude Code a concise pattern for refreshing directory-level CLAUDE.md files.
Debug Code with Root Cause Discipline
Bug fixes often fail when teams guess before they understand the cause. This skill guides Claude, Codex, and Claude Code through evidence-based debugging before implementation.
Route CC-DevFlow Workflows
Teams lose time choosing the next CC-DevFlow command during complex delivery work. This skill maps project status, phase gates, and user intent to the right Claude, Codex, or Claude Code workflow.
Capture Requirements Before Implementation
Teams lose time when vague requests become rushed implementation. This skill structures early brainstorming, compares options, and records decisions for later alignment.
Enforce Test-First Task Completion
Teams can drift from test-first discipline when task files are edited by hand. This skill defines checks that block implementation completion until related tests are complete.
Review Feedback with Technical Rigor
Code review feedback can be incomplete, unclear, or incorrect. This skill helps Claude, Codex, and Claude Code verify feedback before changing code.
Manage Context-Driven Development
by wshobson
AI-assisted projects lose consistency when context lives in scattered notes. This skill provides a structured Conductor workflow for product, stack, workflow, and track artifacts.
Implement Tasks with TDD Workflow
by sickn33
Task plans can drift when implementation, tests, commits, and approvals are handled inconsistently. This skill provides a structured TDD workflow with checkpoints, verification gates, and git tracking.
Build Reliable Test Automation
by 89jobrien
Weak tests slow releases and hide regressions. This skill guides test strategy, TDD, automation, coverage review, and Playwright web checks.
Plan Reliable Tests Across Your Codebase
by 0xDarkMatter
This skill helps teams choose test strategies, mocks, data fixtures, coverage gates, and CI patterns without mixing test scopes. It gives Claude, Codex, and Claude Code structured guidance for unit, integration, end-to-end, TDD, and pipeline testing.
Enforce Waterfall Phase Gates
by Chemiseblanc
AI coding sessions can skip requirements, design, or testing when pressure rises. This skill keeps work inside a strict waterfall flow with documented gates.
Build Python Tests with pytest TDD
by 89jobrien
Python teams often need consistent tests, coverage, and reports without losing TDD discipline. This skill provides pytest-focused workflows, examples, and reporting guidance for Claude, Codex, and Claude Code.