← Back to registry
Quality Advanced by Agent Buildprint
Superpowers Skill Methodology Harness
Make coding agents load relevant skills, design before implementation, use TDD, delegate safely, and block completion without evidence.
methodology
agent-ready 01 Contract spine
02 Implementation phases
03 Validation gates
04 Prompt handoff
Executable packet spine
The files or runtime artifacts a fresh agent must read or produce before claiming progress.
- BOOTSTRAP_AND_ROUTING.md
- BUILDPRINT.md
- CONTRACTS.md
- SKILL_LIBRARY.md
- SOURCE_TRACE.md
- SPEC.md
- SUBAGENT_ORCHESTRATION.md
- README.md
- PLAN.md
- TEST_MATRIX.md
- VALIDATION_TEMPLATE.md
- checks/acceptance.md
Proof gates
Checks that prevent a vague implementation from being reported as complete.
- ✓Skills load before action
- ✓Plans include exact paths/tests
- ✓Review evidence blocks completion
- ✓Transcript evals catch process drift
Risks covered
Failure modes the Buildprint makes visible before an agent can hide them in “done”.
- ⚠Skipping process gates
- ⚠Copying upstream text
- ⚠Claiming official compatibility
What this ships
01
Skill lookup discipline
02
Design approval gate
03
TDD gate
04
Subagent packets
05
Transcript evals
Validation evidence
For newer phase-flow Buildprints, this names replay or outcome evidence. For older registry entries, it lists the included validation checks.
- Skills load before action
- Plans include exact paths/tests
- Review evidence blocks completion
- Transcript evals catch process drift
Copyable agent prompt
Use the Agent Buildprint: Superpowers Skill Methodology Harness.
Bootstrap it with `agb start https://agent-buildprint.com/buildprints/superpowers-skill-methodology-harness/package.json ./my-build` or inspect the GitHub folder. Follow BUILDPRINT.md as the authority spine, read the supporting contracts/checks before implementation, preserve the stated non-goals and safety gates, and do not claim completion until the included validation checks pass.