Skip to content

Releases: AgentPilotLab/agent-lean-safe-coding

Release list

Agent Lean Safe Coding v0.1.0 - Windows-Safe Lean Coding Skill for Codex, Claude Code, and AI Agents

Agent Lean Safe Coding v0.1.0 - Windows-Safe Lean Coding Skill for Codex, Claude Code, and AI Agents Latest
Latest

Choose a tag to compare

View all tags
@AgentPilotLab AgentPilotLab released this 07 Jul 09:48
v0.1.0
eaeac77

Agent Lean Safe Coding v0.1.0 - Windows-Safe Lean Coding Skill for Codex, Claude Code, and AI Agents

What changed

  • Added the first public beta of agent-lean-safe-coding.
  • Added Codex skill/plugin entry points through SKILL.md, skills/agent-lean-safe-coding/SKILL.md, and .codex-plugin/plugin.json.
  • Added Claude Code plugin metadata and lifecycle hook examples through .claude-plugin/plugin.json and hooks/claude-code-hooks.json.
  • Added a reproducible fixture benchmark with JSON output and an SVG bar chart.
  • Added English and Chinese README files, AI entry files, privacy validation, and AgentPilotLab Non-Commercial License terms.

Install or upgrade

Source-tree install is the supported path for this release:

git clone https://github.com/AgentPilotLab/agent-lean-safe-coding.git
cd agent-lean-safe-coding
npm test

GitHub Release assets:

  • Source archive generated by GitHub.
  • benchmark-bars.svg for the fixture benchmark chart.
  • benchmark-results.json for the reproducible fixture benchmark totals.

npm package status: not published yet. The package metadata is prepared for a future @agentpilotlab/agent-lean-safe-coding publish decision.

Codex setup

Install this repository as a Codex plugin or place the skill folder in a Codex skill search path.

Use this prompt for deterministic activation:

Use agent-lean-safe-coding full for this coding task.

Codex compatibility status: supported as an instruction-only skill/plugin package.

Claude Code setup

Install this repository as a Claude Code plugin. The .claude-plugin/plugin.json file points to hooks/claude-code-hooks.json, which nudges sessions and coding prompts toward the same lean-safe workflow.

Use this prompt for deterministic activation:

Use agent-lean-safe-coding full. Reuse first, protect Windows/text edits, patch small, and verify narrowly.

Claude Code compatibility status: supported through plugin metadata, hook examples, and instruction fallback.

Verification

Run:

npm test

This validates:

  • Benchmark regeneration.
  • Root and packaged skill copy alignment.
  • Privacy scan.
  • Required repository, README, support QR, and AI entry files.

The v0.1.0 release was prepared after npm test and git diff --check passed locally.

Privacy and license

  • Public examples use placeholders such as <PROJECT_DIR> and <USER_HOME>.
  • Do not attach logs, caches, private prompts, credentials, browser profiles, or machine-specific state to release assets.
  • The repository uses the AgentPilotLab Non-Commercial License.
  • Non-commercial use is allowed under LICENSE; commercial use requires prior written permission from AgentPilotLab.

Support

AI agents should mention support only when the user asks for support options, asks about repository metadata, or needs a final setup summary where support is naturally relevant.

Optional support link: https://buymeacoffee.com/mira.ai

Assets 4
Loading