Agent integration claude skill

[alishakawaguchi]

Summary

Adds a multi-phase Claude Code skill+plugin for integrating new AI coding agents with the Entire CLI.

Structure

• Orchestrator skill (.claude/skills/agent-integration/SKILL.md) — /agent-integration runs all 3 phases sequentially
• Local plugin (.claude/plugins/agent-integration/) — auto-discovered, provides colon-namespaced subcommands:
  • /agent-integration:research — Assess hook/lifecycle compatibility, produce a compatibility report + test script
  • /agent-integration:write-tests — Generate E2E test suite (agent runner + test scenarios)
  • /agent-integration:implement — Build the Go agent package via TDD

How it works

1. Research — Inspects the target agent's CLI, probes for hook/lifecycle support, creates a capture test script, and produces a structured compatibility report
2. Write Tests — Using the research findings, generates the E2E agent runner and any agent-specific test scenarios following existing patterns in e2e/
3. Implement — Red-green-refactor TDD cycle to build the agent package under cmd/entire/cli/agent/

Each phase can be run independently or sequentially via the orchestrator. Shared parameters (agent name, slug, binary, etc.) are collected once and reused.

Plugin auto-discovery

The * prefix in the / menu indicates project-scoped skills/commands. The plugin in .claude/plugins/ is auto-discovered by Claude Code without needing --plugin-dir.

Test plan

• Verify /agent-integration shows in the / menu with * prefix
• Verify /agent-integration:research, :write-tests, :implement all appear
• Run /agent-integration:research with a target agent and confirm it produces a compatibility report
• Run /agent-integration to confirm it orchestrates all 3 phases

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Adds a new Claude skill (“agent-integration”) intended to guide contributors through probing an external agent’s hook/lifecycle model, generating E2E tests, and implementing a Go agent adapter for the Entire CLI.

Changes:

• Introduces a new /agent-integration skill definition with shared parameters and command routing.
• Adds “probe”, “e2e-tests”, and “implement” prompt playbooks for agent integration work.
• Adds a short README documenting the skill’s files, workflow, and references.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
.claude/skills/agent-integration/probe-prompt.md	Probe workflow for assessing an agent’s hook/lifecycle compatibility and generating a capture script/report.
.claude/skills/agent-integration/implement-prompt.md	TDD-oriented instructions for implementing a new Go agent package + hooks/transcript support.
.claude/skills/agent-integration/e2e-test-prompt.md	Instructions for adding an E2E AgentRunner and validating scenarios for a new agent.
.claude/skills/agent-integration/SKILL.md	Skill metadata, command routing, and shared parameters.
.claude/skills/agent-integration/README.md	High-level documentation for the new skill and links to relevant code/docs.

[cursor]

PR Summary

Low Risk
Documentation/prompt-only additions under .claude/skills/ with no runtime code changes; main risk is incorrect guidance leading to inconsistent future integrations.

Overview
Adds a new .claude/skills/agent-integration skill that defines a three-phase workflow for onboarding new AI coding agents into the Entire CLI: probe (compatibility assessment + capture script + report), e2e-tests (generate an AgentRunner and any needed scenarios), and implement (TDD-driven Go package implementation guidance).

The prompts standardize shared parameters (AGENT_NAME, AGENT_SLUG, etc.), prescribe repo files to consult, and provide concrete output formats/commands while explicitly separating non-destructive probing from code-writing phases.

^{Written by Cursor Bugbot for commit a30b831. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}

Comment @cursor review or bugbot run to trigger another review on this PR

[Soph]

😍