Back to registry
↻ Workflow OS ✓ Validated Updated Not dated

Superpowers Skill Methodology Harness

A blueprint for building an actual coding-agent skill setup harness, not a manager: startup bootstrap, skill activation, methodology gates, subagent execution, code review, branch finishing, and transcript tests.

Package JSON Source: obra/superpowers
What this builds

A reusable agent workflow packet.

A source-backed Buildprint for reconstructing the architecture and behavior contracts behind obra/superpowers as a portable skill-methodology harness.

  • Session-start bootstrap contract
  • Skill metadata and routing model
  • Brainstorming-before-code workflow
  • Spec and implementation-plan artifacts
  • TDD hard gate
  • Systematic debugging gate
Core capabilities

The useful parts the finished build should expose.

01 Session-start bootstrap contract
02 Skill metadata and routing model
03 Brainstorming-before-code workflow
04 Spec and implementation-plan artifacts
05 TDD hard gate
06 Systematic debugging gate
What you need

Local first, live proof explicit.

  • One target agent runtime to support first
  • A skill loading mechanism or plugin adapter
  • A transcript/eval runner
  • A safe subagent/task interface
  • Human policy for exceptions to TDD/design gates
System shape

What kind of artifact this becomes.

Workflow surface

A source-backed Buildprint for reconstructing the architecture and behavior contracts behind obra/superpowers as a portable skill-methodology harness.

Runtime layer

Claude Code, Codex, OpenCode, Gemini, Cursor, Copilot CLI

Build materials

Bootstrap / Skill lookup / Design / Plan

Proof boundary

Source-backed mapping / Actual skill harness

Build scope

Included, required from you, and outside the claim.

Included
  • Session-start bootstrap contract
  • Skill metadata and routing model
  • Brainstorming-before-code workflow
  • Spec and implementation-plan artifacts
Bring yourself
  • One target agent runtime to support first
  • A skill loading mechanism or plugin adapter
  • A transcript/eval runner
  • A safe subagent/task interface
Out of scope
  • Confusing skill files with a working harness
  • Skipping bootstrap injection
  • Skills not auto-triggering
  • Implementation before design approval
Agent handoff

Start from the packet, not the UI.

agb start https://agent-buildprint.com/buildprints/superpowers-skill-methodology-harness/package.json
Agent guide Manifest GitHub files
Key files

The first files an agent should read.

BUILDPRINT.md compatibility bootstrap or package contract
All package files
BOOTSTRAP_AND_ROUTING.md Buildprint package file
BUILDPRINT.md compatibility bootstrap or package contract
checks/acceptance.md acceptance checklist
CONTRACTS.md legacy interface/data contracts, when present
PLAN.md legacy execution index, when present
proof/package.json offline proof artifact
proof/src/eval.js offline proof artifact
proof/src/harness.js offline proof artifact
proof/src/run-evals.js offline proof artifact
proof/src/skills.js offline proof artifact
proof/src/subagents.js offline proof artifact
proof/src/transcript.js offline proof artifact
proof/tests/harness.test.js offline proof artifact
publication.json machine-readable mirror
README.md human overview, non-authoritative
SKILL_LIBRARY.md Buildprint package file
SOURCE_TRACE.md Buildprint package file
SPEC.md legacy behavior requirements, when present
SUBAGENT_ORCHESTRATION.md Buildprint package file
SYSTEM_MAP.md Buildprint package file
TEST_MATRIX.md legacy risk-to-test alignment, when present
VALIDATION_REPORT.md Buildprint package file
VALIDATION_TEMPLATE.md legacy completion report template, when present
WORKFLOW.md Buildprint package file