ooooooooooooooooooop/agent-tools
Instructions file CodexOpenCode
Repository instructions for an agent-tools project, which packages coding-agent skills, MCP tools, plugins, and quality checks. They describe the directory layers, allowed changes, and required validation.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A placeholder skill description written in Chinese that explains what a skill does and when it should start.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
An operating guide for coordinating multiple coding agents through a manager-and-worker process.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A pre-change checklist for clarifying the scope, risks, assumptions, affected files, and acceptance criteria of a requested modification.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A set of checkpoints for keeping long, multi-step agent tasks aligned with their original goal and evidence.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A tool for packaging and restoring DeepSeek Harness user settings across devices. DeepSeek Harness is a coding-agent environment; the package contains preferences and model-routing settings, not credentials or runtime data.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A read-only and approval-based process for restoring registered Codex skills from a source repository to another device or user-level skills folder.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A process for turning recorded audit problems into reviewable proposals for changes to rules, skills, plugins, or routing. It uses evidence and requires approval before changes are applied.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A set of rules for carrying out long, multi-tool coding-agent sessions without wasting work or losing track of progress. It covers waiting for background tasks, handling failed checks, assigning subagents, saving results, and closing each work round cleanly.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A tool for finding past DSH conversations by title or text and reading their content or usage statistics. DSH is the system referenced by the skill for storing agent sessions.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A router that chooses how deeply a coding-agent task should be handled based on its size and risk. It distinguishes small edits, read-only reviews, medium changes, large projects, and high-risk actions such as installing tools or changing global settings.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A Chinese- and English-language rewriting skill for making messages, emails, comments, explanations, and social posts sound natural. It preserves the original facts, meaning, tone, and uncertainty.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A read-only health review for a finished Personal AI or DSH infrastructure setup. It checks current or previously recorded evidence and labels areas as healthy, degraded, blocked, or unknown.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A workflow for keeping an AI setup consistent across multiple computers. It decides whether to download, upload, merge, review, or block changes, while leaving conflicts and unsaved changes for a person.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A research workflow for questions that can be tested with evidence, such as comparing models or software. It defines the question, baseline, candidates, data, measurements, and decision rules before running or resuming an experiment.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A structured discussion method for difficult decisions with competing constraints or uncertain information. It compares perspectives from selected real experts, a general field expert, and an abstract all-knowing viewpoint, while marking guesses separately from facts.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A quality review for a coding-agent Skill, meaning a package of instructions and supporting files. It examines when the Skill should run, how it works, what it returns, how it is tested, and what safety limits it has.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A maintenance guide for checking, auditing, and synchronising a Codex Skill repository. It covers package structure, metadata, release boundaries, and differences between the source and installed folders.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A set of rules for assigning coding work to subagents, which are separate AI workers. It separates discovery, planning, implementation, and validation, and limits what each worker may read and write.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A structured problem-solving and improvement workflow for systems that are slow, costly, error-prone, or repeatedly failing. It starts with measured evidence, then traces root causes and checks whether changes worked.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A task-routing guide that chooses how deeply to handle a request based on its size and risk. It covers code, configuration, documentation, workflows, installations, and other changes.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A task-management workflow for genuinely large, multi-stage jobs that need saved plans, checkpoints, recovery information, and acceptance tracking. A checkpoint is a recorded progress marker used to resume work safely.
ooooooooooooooooooop/agent-tools
Skill Claude CodeCodex
A fixed workflow that turns three people’s daily Excel reports into one weekly Chinese work-summary spreadsheet, using Shanghai time and China’s actual working days.