feat: close superpowers gaps — anti-rationalization, status protocol, multi-platform#18
Open
bdfinst wants to merge 15 commits into
Open
feat: close superpowers gaps — anti-rationalization, status protocol, multi-platform#18bdfinst wants to merge 15 commits into
bdfinst wants to merge 15 commits into
Conversation
Competitive analysis comparing agentic-dev-team against obra/superpowers identified gaps in anti-rationalization techniques, code review reception, skill authoring depth, debugging references, subagent status codes, TDD depth, and worktree setup. Specs cover 7 core implementation slices (A-G) plus 3 platform support slices (H1-H3: Windows hooks, Gemini CLI, OpenAI Codex). Multi-platform research document recommends graceful degradation with platform-specific manifests reusing existing skills. Removes completed/obsolete spec files from prior work.
18-step plan covering 10 slices (A-G core improvements, H1-H3 platform support). All four plan review personas approved. Acceptance critic blockers resolved: binary-verifiable criteria, parser validation commands, deterministic DONE_WITH_CONCERNS branches.
Slice A: Anti-rationalization knowledge file + quality gate references Slice B: Receiving-code-review skill with verification gates Slice C: Pressure testing + CSO checklist in skill authoring guide Slice D: Debugging supporting files (root-cause tracing, condition-based waiting, find-polluter bisection) Slice H1: Windows hooks shim (run-hook.cmd), install.ps1 prereq checker, TMPDIR fixes in tdd-guard.sh and version-check.sh
Steps 7-8: Create implementer, spec-reviewer, and quality-reviewer prompt templates with full behavioral content and status blocks Step 9: Add status field to all 4 plan review templates Step 10: Add Subagent Status Protocol section to orchestrator Step 11: Add status handling to build command step 4
Create testing anti-patterns reference file. Add cross-reference from TDD catch-all line to anti-rationalization knowledge. Add testing anti-patterns loading note to implementer prompt.
…ewer overlap Slice G: Create worktree-setup knowledge file with language detection table. Add Worktree Setup section to implementer prompt. Reference worktree setup in orchestrator Phase 3. Fix: Rewrite quality-reviewer from a direct reviewer to a dispatcher that selects and runs specialized review agents based on what changed. Eliminates overlap with naming-review, structure-review, complexity-review, test-review, and security-review agents.
Slice H2: Gemini CLI extension manifest, GEMINI.md context file, 5 TOML commands, hooks-gemini.json placeholder Slice H3: AGENTS.md for Codex, .codex/config.toml, .codex/hooks.json placeholder, CODEX-INSTALL.md setup guide Both platforms reuse existing skills/ and knowledge/ without forking. Capability limitations documented in each platform context file.
Skills: 31→32 (+receiving-code-review) Knowledge files: 6→8 (+anti-rationalization, +worktree-setup) Prompt templates: 8→7 (removed stale plan-reviewer.md reference) Supporting reference files: +4 (testing-anti-patterns, root-cause-tracing, condition-based-waiting, find-polluter) Skills-by-phase: add receiving-code-review to Review phase Add subagent status protocol reference to Multi-Agent Collaboration
- docs/skills.md: add 4 missing skills (receiving-code-review, static-analysis-integration, domain-analysis, js-project-init), remove stale plan-reviewer reference, update quality-reviewer description to "dispatcher" - docs/agent_info.md: add prompt templates section for implementer, spec-reviewer, and quality-reviewer - CLAUDE.md: skill count 32→33 (+js-project-init) - GEMINI.md + AGENTS.md: skill count 31→33 - agent-registry.md: add js-project-init, remove stale plan-reviewer
The team-agents.svg was showing 10 agents but there are 11 — ADR Author was missing. Added to center column below Product Manager.
Rename agent and command files, update all 14 references across the plugin (CLAUDE.md, GEMINI.md, AGENTS.md, agent-registry, context-loading, ci-debugging, performance-benchmark, security-engineer, benchmark command, agent_info docs, team-agents diagram, GETTING-STARTED guide).
New knowledge file: contract-testing.md defines the two-layer validation pattern — test doubles in the deterministic pipeline, integration tests to validate the doubles against real dependencies. TDD skill: replace "mocks only when unavoidable" with "validated test doubles at architectural boundaries, real code for internal modules." testing-anti-patterns: rewrite #5 from "integration tests as afterthought" to "unvalidated test doubles" — the real anti-pattern is doubles without integration tests that verify them, not the use of doubles itself. Source: https://beyond.minimumcd.org/docs/testing/contract/index.html.md
Add all Claude Code tools to the allow list so the plugin operates without permission prompts. Narrow deny rules from broad patterns (rm *, git push *) to only truly irreversible operations (rm -rf, force-push).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Closes gaps identified in competitive analysis against obra/superpowers. 10 slices:
Also rewrote quality-reviewer.md from direct reviewer to dispatcher, eliminating overlap with naming-review, structure-review, complexity-review, test-review, and security-review.
Quality Gate
Test Plan
gemini-extension.jsonand.codex/config.tomlparse correctly🤖 Generated with Claude Code