Execute external AI agent CLIs via Bash for delegated coding tasks.
npx skills add https://github.com/xuiltul/animaworks --skill subagent-cli
Please help me install this Agent Skill. Skill: subagent-cli Source: https://github.com/xuiltul/animaworks/tree/main/templates/ja/common_skills/subagent-cli Command: npx skills add https://github.com/xuiltul/animaworks --skill subagent-cli
This Skill provides a safe, repeatable way to delegate complex, multi-file coding, review, and refactoring tasks to external AI agent CLIs via Bash while retaining the host agent's identity, memory, and decision logic.
Run a background codex exec targeting the project workspace with --full-auto and --ephemeral, saving the agent output to a result file and using a timeout to prevent runaway execution.
High-intent search queries and answers about installing and using this skill.
You can delegate multi-file refactoring tasks by executing external AI agent CLIs noninteractively via Bash, specifying a workspace directory, applying sandboxing flags, and capturing outputs to files using a codex to cursor to claude fallback strategy.
Yes, long-running code generation tasks can run in the background using nohup and timeout commands, allowing the host agent to continue operating asynchronously while capturing agent outputs to designated result files.
The fallback strategy for automated code review prefers codex exec first, then falls back to cursor-agent -p, and finally uses claude -p with cost limits and sandboxing flags to ensure task completion.
Yes, delegating complex coding to external AI agents requires Bash execution permission across Mode S/A/B, with optional Mode C access for certain CLIs, to safely execute workspace-targeted operations.
To audit changes after multi-file modifications, you should perform post-run git diff checks on the workspace directory, leveraging the recommended git-based workflows to track and review all changes made by the subagent.
Using ephemeral sessions with --full-auto flags enhances safety by preventing state persistence, but you must enforce explicit workspace directories and prompt hygiene to avoid unintended file modifications during background execution.
jal-coDefine and configure OpenCD agents with permissions, tools, and mode settings.
d-oitExecute safe commands and capture stdout, stderr, and exit codes.
HenquewlExecute software development tasks from predefined implementation plans with verification checkpoints.
ever-oliDelegate coding tasks to a multi-model AI agent with an integrated judge.
