vadimcomanescu/codex-skills
Overview
This skill transforms workflows to use persistent Manus-style markdown files for planning, progress tracking, and knowledge storage. It standardizes how you start complex tasks, record findings, and produce final deliverables using simple file conventions. The approach reduces reliance on ephemeral chat context and creates an auditable project trail.
How this skill works
When enabled, the skill prompts creation of a task plan file (e.g., task_plan.md) before starting any complex or multi-step work. It directs research and raw notes into a separate notes.md, guides you to write deliverables in dedicated output files, and recommends updating the plan file after each phase to record status and errors. Files serve as persistent working memory that the agent reads and writes to keep context across sessions.
When to use it
- Starting a complex or multi-step project that needs tracking and checkpoints
- Conducting research where findings should be stored outside chat for later reference
- Coordinating work that requires repeatable status updates and error logs
- Preparing deliverables that must be drafted, reviewed, and versioned
- When you want an auditable trail of planning decisions and progress
Best practices
- Create task_plan.md before initiating substantive work and list goals, milestones, and owners
- Keep raw findings and source material in notes.md rather than chat to avoid context loss
- Write final outputs in a dedicated deliverable file and link back to the plan
- Update task_plan.md after each phase with status, next steps, and any errors encountered
- Store large data or examples in reference files (e.g., references/examples.md) and cite them in the plan
Example use cases
- Planning a multi-week product feature with milestones, dependencies, and progress logs
- Running a research sprint where literature, experiments, and summaries live in notes.md
- Managing a complex bugfix process with reproducible steps and an error log in the plan file
- Preparing a report: outline in task_plan.md, research in notes.md, and final text in deliverable.md
- Handing off work: the plan file documents decisions and next steps for the incoming contributor
FAQ
Start with task_plan.md for goals and milestones, and notes.md for research and raw findings.
How often should I update the plan file?
Update task_plan.md after each meaningful phase or shift—daily for active sprints, or after each major milestone.
18 skills
This skill helps you plan and organize complex tasks using persistent Markdown files for planning, tracking, and knowledge storage.
This skill helps you design and document scalable architectures with ADRs, diagrams, and clear interfaces to inform decisions and guide implementation.
This skill helps you design reliable data pipelines by defining contracts, quality checks, and observability to ensure predictable, recoverable ETL workflows.
This skill helps you build and deploy continuous autonomous agents with a visual flow editor and modular blocks for rapid automation.
This skill accelerates security incident triage and remediation by guiding containment, eradication, recovery, and postmortem learning.
This skill helps you build reliable API integrations with strong auth, error handling, retries, and boundary payload transformations.
This skill guides you through test-driven development, writing failing tests first and making minimal changes to drive design and prevent regressions.
This skill helps you start isolated feature work by creating git worktrees with safe paths, baseline checks, and clean status verification.
This skill helps you diagnose errors methodically, reproduce issues, and validate safe, minimal fixes with clear verification steps.
This skill helps you create, edit, and analyze Excel workbooks using formulas, templates, and data visualization while preserving existing structures.
This skill helps design, evaluate, and deploy computer vision systems by guiding task definition, dataset strategy, and performance-oriented workflows.
This skill helps you design clear, scalable information architectures for apps and docs, improving navigation, labeling, and user wayfinding.
This skill enables time series machine learning tasks such as forecasting, classification, anomaly detection, and clustering using aeon with scikit-learn
This skill helps Senior Product Owners generate INVEST-compliant user stories, plan sprints, and manage backlog with velocity tracking.
This skill helps translate asks into measurable metrics and experiments, design robust analyses, and communicate assumptions and caveats clearly.
This skill helps you design risk-based QA plans, optimize test layer choices, and triage flaky tests to ensure reliable releases.
This skill helps you generate music and sound effects from text descriptions using AudioCraft, enabling melody conditioning, style transfer, and stereo output.
This skill helps you leverage LlamaIndex to build retrieval-augmented generation apps, ingest data, and query private knowledge efficiently.
