oil-oil/codex-team-mode / skills/team-mode
team-mode
VerifiedUse when a task may benefit from a bounded subagent for implementation, large-codebase discovery at task start, independent review, requested or pre-commit code simplification, or expert work on complex decisions, modeling, automation, or repeated failures. The main agent chooses whether to delegate and accepts the result. Do not use for simple questions or short work with no meaningful review or delegation.
| File | Kind | Size |
|---|---|---|
| README.md | markdown | 1.8 KB |
| SKILL.md | skill md | 5.2 KB |
| agents/openai.yaml | config | 410 B |
| references/custom-agents.md | markdown | 4.1 KB |
| references/evaluation.md | markdown | 2.3 KB |
| references/explore.md | markdown | 654 B |
| references/interactive-testing.md | markdown | 1.4 KB |
| references/simplify.md | markdown | 2.9 KB |
| scripts/current_model.pyexecutable | script | 3.4 KB |
| scripts/usage_by_model.pyexecutable | script | 36.0 KB |
references/simplify.md
View on GitHub# Simplify
Wait until a coherent change is settled. Review its changed code and related documentation, not the whole repository. Honor any narrower scope set by the parent or user. Read applicable project instructions such as `AGENTS.md` and follow the repository's conventions. If there is no substantive code change and no explicit request to simplify documentation, report that there is nothing to simplify rather than widening the scope.
Look for complexity that makes the result harder to understand or change: unnecessary gates and special cases, fallback chains that obscure intended behavior or hide failures, stale documentation that would mislead the next agent or maintainer, oversized files, and responsibilities that would be clearer as components or functions. Notice duplication and needless abstraction too. These are leads, not required findings; keep useful safeguards and avoid splitting code solely to make files shorter.
Choose the number of fresh `Reviewer` agents by the number of independent review lenses:
- One Reviewer is the default for a narrow diff or a single main concern. It can assess reuse, clarity, and efficiency together when those are all relevant.
- Use two or three Reviewers when a broader or riskier change has distinct lenses that can be reviewed independently. Examples include reuse of existing helpers, code structure and naming, and runtime or resource efficiency. Give each Reviewer the same resolved source scope, a different lens, and a concrete return format.
- Add more only when additional independent subsystems or risk areas justify separate work. Do not duplicate general reviews or split work by file count alone.
Ask each Reviewer to identify concrete, behavior-preserving simplifications with locations, reasons, and a proposed change. Reviewers report findings and leave source files unchanged in this pass. The installed Reviewer profile may still have workspace-write permission, so this is an instruction boundary, not a sandbox guarantee. Prefer returning findings inline; if a lasting report is needed, assign a unique report path outside the changed source diff.
The main agent selects findings and applies worthwhile changes or assigns an `Executor` with explicit write ownership. Preserve observable outputs, errors, side effects, ordering, trust-boundary validation, data-loss protection, security checks, and accessibility behavior. If equivalence is uncertain, skip the proposed simplification.
After edits, run the validation required by the task and project, and report which checks actually ran and their results. If any reviewed file changes after the Simplify review, dispatch one fresh `Reviewer` to inspect the resulting diff for correctness and omissions. If the reviewed files do not change, the original Reviewer may also serve as the final reviewer only when its brief explicitly included correctness and omission checks. Otherwise, dispatch a fresh Reviewer for the final check.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |