skunkworks-sidecar

Create a quarantined git worktree for isolated proof-of-concept work on high-risk items.

Updated Jun 4, 2026
One-click install
npx skills add https://github.com/Kohlex/autopilot-plugin --skill skunkworks-sidecar
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skunkworks-sidecar
Source: https://github.com/Kohlex/autopilot-plugin/tree/main/skills/skunkworks-sidecar
Command: npx skills add https://github.com/Kohlex/autopilot-plugin --skill skunkworks-sidecar

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The skunkworks-sidecar provides a quarantine-enabled mechanism to de-risk high-risk items by creating a flat-sibling git worktree for isolation and redirecting all GSD sinks under a quarantined planning subtree, preserving the main product's LEARNINGS and STATE intact.

Core Features & Use Cases

  • Flat-sibling worktree-based de-risking: creates a dedicated worktree for high-risk items off the main branch.
  • Sink isolation and provenance: redirects LEARNINGS/STATE/claude-mem/checkpoints/spikes into the quarantine subtree to avoid contaminating main planning state.
  • Proven learnings extraction and injection: on success, produces a skunkworks-proven@<item> payload for /peer-review via B11; on failure, records a single risk entry in a risk ledger.
  • Guardrails: HARD_STOP for sibling products to prevent out-of-fence work on others and quarantined scaffolding with /orient guard to prevent mis-reading as live work.

Quick Start

Run the skunkworks-sidecar workflow to isolate a high-risk item and quarantine its planning data.

Frequently Asked Questions about skunkworks-sidecar

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I isolate high-risk proof-of-concept work from my main product state?

You isolate high-risk proof-of-concept work by spinning a flat-sibling git worktree that redirects all planning sinks into a quarantine subtree, keeping your main product's LEARNINGS and STATE intact.

What is a git worktree used for in de-risking product development?

In de-risking product development, a git worktree provides a flat-sibling isolation environment for high-risk items, ensuring that experimental planning data and checkpoints remain quarantined away from the live branch.

How do I extract proven learnings from a quarantined skunkworks POC?

You extract proven learnings from a successful quarantined skunkworks POC by generating a skunkworks-proven payload for peer review, ensuring validated knowledge is safely injected back into the main workflow.

What happens to the risk ledger when a high-risk POC fails?

When a high-risk POC fails, the system records a single risk entry in a risk ledger, ensuring that the failure is documented without contaminating the main product's planning state.

Does this de-risking workflow have guardrails to prevent accidental out-of-fence work?

Yes, the de-risking workflow enforces HARD_STOP guardrails for sibling products and uses quarantined scaffolding with orientation guards to prevent users from misreading experimental work as live work.

When should I use a quarantine sidecar for git worktree automation?

You should use a quarantine sidecar when front-load risk flags trigger POC work that requires isolated learnings, controlled merge-back, and strict fence discipline to avoid contaminating main planning state.