Name: Evolve
Author: klintravis

Prompt file imported from klintravis/CopilotCustomizer (.github/prompts/Evolve.prompt.md). Fill in {{task}}, {{scope}}, {{depth}}, {{autoExecute}} before use. Copyright stays with the author.

Evolve (v2.0)

Task Intent

Quick entry point for common CopilotCustomizer toolkit improvement workflows. Routes through the CopilotCustomizer orchestrator to the Evolve specialist subagent with pre-configured task context for streamlined self-improvement.

Variable Block

Task [OPTIONAL]: {{task}} Scope [OPTIONAL]: {{scope}} Depth [OPTIONAL]: {{depth}} Auto-Execute [OPTIONAL]: {{autoExecute}}

Task Descriptions

Task	Description	Typical Duration
`monitor-releases`	Check VS Code Copilot docs for new releases and features	2–3 minutes
`review-docs`	Analyze README/HOW-TO/QUICKSTART for quality and currency	3–5 minutes
`audit-assets`	Run self-audit on toolkit's own `.github/` assets	3–5 minutes
`optimize`	Identify token efficiency and performance improvements	2–4 minutes
`feature-request`	Orchestrate implementation of a toolkit feature or correction	5–15 minutes
`self-validate`	Pre-release validation of all toolkit assets	4–6 minutes

Scope Filters

Scope	Assets Included
`all`	Agents, instructions, prompts, skills, documentation
`agents`	`.github/agents/*.agent.md`
`instructions`	`.github/instructions/*.instructions.md`
`prompts`	`.github/prompts/*.prompt.md`
`skills`	`.github/skills/*/SKILL.md`
`docs`	README.md, QUICKSTART.md, HOW-TO.md, EXAMPLES.md, AGENTS.md, docs/ARCHITECTURE.md

Depth Modes

Depth	Analysis Level	Output Characteristics
`quick-check`	Surface scan	Top 3 findings + action items
`standard`	Full audit	Complete findings table + prioritized roadmap
`comprehensive`	Deep analysis	+ regression check, cross-reference verification, performance metrics

Workflow Phases

Phase 1: Analysis (Auto) — Execute task-specific analysis workflow per TASK and SCOPE Phase 2: Recommendations (Auto) — Generate prioritized improvement recommendations at DEPTH level Phase 3: Approval Gate (User) — Review recommendations and approve actions Phase 4: Execution (Conditional) — If AUTO_EXECUTE=true for low-risk items, or user approves, implement changes Phase 5: Verification (Auto) — Validate all changes via Verifier Phase 6: Documentation (Auto) — Generate summary report with changelog entries Phase 7: Plan Archival (Auto) — Persist approved plan to docs/plans/ for audit trail

Orchestration Chain

Evolve → CopilotCustomizer orchestrator → Evolve → [Editor] → [Verifier] → Complete

Example Invocations

Default audit (most common):

/Evolve

Runs asset audit with standard depth across all toolkit assets.

Quick release check:

/Evolve task: "monitor-releases", depth: "quick-check"

Fast scan of VS Code Copilot docs for new features.

Comprehensive doc review:

/Evolve task: "review-docs", scope: "docs", depth: "comprehensive"

Deep review of all toolkit documentation.

Optimize agents only:

/Evolve task: "optimize", scope: "agents", autoExecute: "false"

Analyze agent token usage with manual approval before changes.

Pre-release validation:

/Evolve task: "self-validate", depth: "comprehensive"

Full validation suite before a toolkit release.

Implement a feature:

/Evolve task: "feature-request"

Then describe the feature in the chat. The orchestrator routes to Evolve, which will classify, plan, and implement the feature.

Refinement Commands

Command	Action
`refine: task`	Switch to a different task type
`refine: scope`	Narrow or expand asset scope
`refine: depth`	Adjust analysis thoroughness
`refine: execute`	Proceed with approved recommendations

Framework References

Paired Agent: Evolve.agent.md Paired Instructions: ToolkitOps.instructions.md Framework: Framework.instructions.md

Generated following CopilotCustomizer prompt generation standards

## Evolve (v2.0) ### Task Intent Quick entry point for common CopilotCustomizer toolkit improvement workflows. Routes through the CopilotCustomizer orchestrator to the Evolve specialist subagent with pre-configured task context for streamlined self-improvement. ### Variable Block --- **Task** [OPTIONAL]: {{task}} **Scope** [OPTIONAL]: {{scope}} **Depth** [OPTIONAL]: {{depth}} **Auto-Execute** [OPTIONAL]: {{autoExecute}} --- ### Task Descriptions | Task | Description | Typical Duration | |------|-------------|------------------| | `monitor-releases` | Check VS Code Copilot docs for new releases and features | 2–3 minutes | | `review-docs` | Analyze README/HOW-TO/QUICKSTART for quality and currency | 3–5 minutes | | `audit-assets` | Run self-audit on toolkit's own `.github/` assets | 3–5 minutes | | `optimize` | Identify token efficiency and performance improvements | 2–4 minutes | | `feature-request` | Orchestrate implementation of a toolkit feature or correction | 5–15 minutes | | `self-validate` | Pre-release validation of all toolkit assets | 4–6 minutes | ### Scope Filters | Scope | Assets Included | |-------|----------------| | `all` | Agents, instructions, prompts, skills, documentation | | `agents` | `.github/agents/*.agent.md` | | `instructions` | `.github/instructions/*.instructions.md` | | `prompts` | `.github/prompts/*.prompt.md` | | `skills` | `.github/skills/*/SKILL.md` | | `docs` | README.md, QUICKSTART.md, HOW-TO.md, EXAMPLES.md, AGENTS.md, docs/ARCHITECTURE.md | ### Depth Modes | Depth | Analysis Level | Output Characteristics | |-------|---------------|------------------------| | `quick-check` | Surface scan | Top 3 findings + action items | | `standard` | Full audit | Complete findings table + prioritized roadmap | | `comprehensive` | Deep analysis | + regression check, cross-reference verification, performance metrics | ### Workflow Phases **Phase 1: Analysis** (Auto) — Execute task-specific analysis workflow per TASK and SCOPE **Phase 2: Recommendations** (Auto) — Generate prioritized improvement recommendations at DEPTH level **Phase 3: Approval Gate** (User) — Review recommendations and approve actions **Phase 4: Execution** (Conditional) — If AUTO_EXECUTE=true for low-risk items, or user approves, implement changes **Phase 5: Verification** (Auto) — Validate all changes via Verifier **Phase 6: Documentation** (Auto) — Generate summary report with changelog entries **Phase 7: Plan Archival** (Auto) — Persist approved plan to `docs/plans/` for audit trail ### Orchestration Chain ``` Evolve → CopilotCustomizer orchestrator → Evolve → [Editor] → [Verifier] → Complete ``` ### Example Invocations **Default audit** (most common): ``` /Evolve ``` Runs asset audit with standard depth across all toolkit assets. **Quick release check**: ``` /Evolve task: "monitor-releases", depth: "quick-check" ``` Fast scan of VS Code Copilot docs for new features. **Comprehensive doc review**: ``` /Evolve task: "review-docs", scope: "docs", depth: "comprehensive" ``` Deep review of all toolkit documentation. **Optimize agents only**: ``` /Evolve task: "optimize", scope: "agents", autoExecute: "false" ``` Analyze agent token usage with manual approval before changes. **Pre-release validation**: ``` /Evolve task: "self-validate", depth: "comprehensive" ``` Full validation suite before a toolkit release. **Implement a feature**: ``` /Evolve task: "feature-request" ``` Then describe the feature in the chat. The orchestrator routes to Evolve, which will classify, plan, and implement the feature. ### Refinement Commands | Command | Action | |---------|--------| | `refine: task` | Switch to a different task type | | `refine: scope` | Narrow or expand asset scope | | `refine: depth` | Adjust analysis thoroughness | | `refine: execute` | Proceed with approved recommendations | ### Framework References *Paired Agent: [Evolve.agent.md](../agents/Evolve.agent.md)* *Paired Instructions: [ToolkitOps.instructions.md](../instructions/ToolkitOps.instructions.md)* *Framework: [Framework.instructions.md](../instructions/Framework.instructions.md)* --- *Generated following CopilotCustomizer prompt generation standards*

{ "prompt": "## Evolve (v2.0)\n\n### Task Intent\nQuick entry point for common CopilotCustomizer toolkit improvement workflows. Routes through the CopilotCustomizer orchestrator to the Evolve specialist subagent with pre-configured task context for streamlined self-improvement.\n\n### Variable Block\n---\n**Task** [OPTIONAL]: {{task}}\n**Scope** [OPTIONAL]: {{scope}}\n**Depth** [OPTIONAL]: {{depth}}\n**Auto-Execute** [OPTIONAL]: {{autoExecute}}\n---\n\n### Task Descriptions\n\n| Task | Description | Typical Duration |\n|------|-------------|------------------|\n| `monitor-releases` | Check VS Code Copilot docs for new releases and features | 2–3 minutes |\n| `review-docs` | Analyze README/HOW-TO/QUICKSTART for quality and currency | 3–5 minutes |\n| `audit-assets` | Run self-audit on toolkit's own `.github/` assets | 3–5 minutes |\n| `optimize` | Identify token efficiency and performance improvements | 2–4 minutes |\n| `feature-request` | Orchestrate implementation of a toolkit feature or correction | 5–15 minutes |\n| `self-validate` | Pre-release validation of all toolkit assets | 4–6 minutes |\n\n### Scope Filters\n\n| Scope | Assets Included |\n|-------|----------------|\n| `all` | Agents, instructions, prompts, skills, documentation |\n| `agents` | `.github/agents/*.agent.md` |\n| `instructions` | `.github/instructions/*.instructions.md` |\n| `prompts` | `.github/prompts/*.prompt.md` |\n| `skills` | `.github/skills/*/SKILL.md` |\n| `docs` | README.md, QUICKSTART.md, HOW-TO.md, EXAMPLES.md, AGENTS.md, docs/ARCHITECTURE.md |\n\n### Depth Modes\n\n| Depth | Analysis Level | Output Characteristics |\n|-------|---------------|------------------------|\n| `quick-check` | Surface scan | Top 3 findings + action items |\n| `standard` | Full audit | Complete findings table + prioritized roadmap |\n| `comprehensive` | Deep analysis | + regression check, cross-reference verification, performance metrics |\n\n### Workflow Phases\n\n**Phase 1: Analysis** (Auto) — Execute task-specific analysis workflow per TASK and SCOPE\n**Phase 2: Recommendations** (Auto) — Generate prioritized improvement recommendations at DEPTH level\n**Phase 3: Approval Gate** (User) — Review recommendations and approve actions\n**Phase 4: Execution** (Conditional) — If AUTO_EXECUTE=true for low-risk items, or user approves, implement changes\n**Phase 5: Verification** (Auto) — Validate all changes via Verifier\n**Phase 6: Documentation** (Auto) — Generate summary report with changelog entries\n**Phase 7: Plan Archival** (Auto) — Persist approved plan to `docs/plans/` for audit trail\n\n### Orchestration Chain\n```\nEvolve → CopilotCustomizer orchestrator → Evolve → [Editor] → [Verifier] → Complete\n```\n\n### Example Invocations\n\n**Default audit** (most common):\n```\n/Evolve\n```\nRuns asset audit with standard depth across all toolkit assets.\n\n**Quick release check**:\n```\n/Evolve task: \"monitor-releases\", depth: \"quick-check\"\n```\nFast scan of VS Code Copilot docs for new features.\n\n**Comprehensive doc review**:\n```\n/Evolve task: \"review-docs\", scope: \"docs\", depth: \"comprehensive\"\n```\nDeep review of all toolkit documentation.\n\n**Optimize agents only**:\n```\n/Evolve task: \"optimize\", scope: \"agents\", autoExecute: \"false\"\n```\nAnalyze agent token usage with manual approval before changes.\n\n**Pre-release validation**:\n```\n/Evolve task: \"self-validate\", depth: \"comprehensive\"\n```\nFull validation suite before a toolkit release.\n\n**Implement a feature**:\n```\n/Evolve task: \"feature-request\"\n```\nThen describe the feature in the chat. The orchestrator routes to Evolve, which will classify, plan, and implement the feature.\n\n### Refinement Commands\n\n| Command | Action |\n|---------|--------|\n| `refine: task` | Switch to a different task type |\n| `refine: scope` | Narrow or expand asset scope |\n| `refine: depth` | Adjust analysis thoroughness |\n| `refine: execute` | Proceed with approved recommendations |\n\n### Framework References\n*Paired Agent: [Evolve.agent.md](../agents/Evolve.agent.md)*\n*Paired Instructions: [ToolkitOps.instructions.md](../instructions/ToolkitOps.instructions.md)*\n*Framework: [Framework.instructions.md](../instructions/Framework.instructions.md)*\n\n---\n\n*Generated following CopilotCustomizer prompt generation standards*", "variables": [ "task", "scope", "depth", "autoExecute" ], "notes": "Quick entry point for CopilotCustomizer toolkit improvement tasks with smart defaults Tools: execute, read, edit, search, web, agent, todo. Arguments: All inputs optional — runs with intelligent defaults.", "metadata": { "source": { "provider": "github-prompts", "repository": "https://github.com/klintravis/CopilotCustomizer", "path": ".github/prompts/Evolve.prompt.md", "ref": "b2af066fbb4b2c8fa71b0596280eafc08b4b2268", "url": "https://github.com/klintravis/CopilotCustomizer/blob/b2af066fbb4b2c8fa71b0596280eafc08b4b2268/.github/prompts/Evolve.prompt.md", "key": "klintravis/CopilotCustomizer/.github/prompts/Evolve.prompt.md" } }, "model_hints": [ "Claude Sonnet 4.5 (copilot)" ] }

Evolve

About