Superpowers Wiki
The /superpowers/ area tracks implementation-focused engineering work: concrete plans, design
decisions, and verification gates used to land non-trivial changes safely.
What Superpowers Covers
Superpowers content is centered on execution artifacts, including:
- scoped implementation plans with ordered tasks
- design decisions tied to specific issue clusters
- explicit backward-compatibility notes
- test strategy and regression coverage per change set
- rollout order to reduce integration risk
- verification gates (
ruff,pytest, and CI expectations)
This section is not product marketing, not high-level onboarding, and not a replacement for the
core docs in /wiki/.
Current Focus
The active published workstream is Code Review Batch 1 (dated 2026-05-18), which resolves four design-bearing clusters:
- Cluster C: remove unsafe article-stripping from context compression (#32)
- Cluster A: treat Azure endpoint/deployment/model metadata as sensitive and redact serialized output (#31, #35)
- Cluster D: deep-merge
opencode.jsonpolicy sections and stage.newupdates (#39) - Cluster B: harden team sync trust boundaries and clone/pull behavior (#30, #36, #43)
Execution order is intentionally C -> A -> D -> B to land the smallest/lowest-risk work first
and the team.py rewrite last on a known-good base.
Recommended Reading
- Code Review Batch 1 - Design Decisions: rationale, mechanisms, tests, compatibility, and out-of-scope boundaries.
- Code Review Batch 1 - Implementation Plan: task-by-task execution checklist, file-level changes, and commit gates.
Operating Posture
Superpowers work is intentionally strict:
- every meaningful behavior change gets a regression test
- risky writes use reviewable staging (
.new) instead of silent overwrite - trust boundaries are explicit for third-party/team-sourced content
- each cluster should be committable independently with green verification
- changes prioritize deterministic, local-first behavior over cleverness