ethanplusai/codex-on-crack / plugins/codex-on-crack/skills/crack
crack
UnverifiedLead substantial Codex builds with one configured implementation worker and lead acceptance. Use when the user requests codex-on-crack, names a model to delegate work to, asks to run subagents through Claude, or requests a lead/worker split. Keep host-only work direct; honor explicit external model choices. Workers must not invoke this orchestrator.
templates/task-brief.md
View on GitHub# <Task ID>: <Reviewable deliverable>
## Assignment
Role agent: crack_<role> (rung: primary | fallback)
Phase: <phase ID>
Workspace: <exact path you verified: the main workspace, or .crack/worktrees/<task ID>>
Baseline: <branch/commit plus pre-existing changes; a worktree does not contain uncommitted main-workspace changes>
Dependencies: <accepted task IDs and the concrete outputs present in this workspace>
Report path: <unique task-owned path>
## Goal and non-goals
<The whole coherent bundle, not one microscopic step.>
## Done when
- <Observable behavior and its acceptance ID.>
- <Relevant failure boundary or state/ownership invariant.>
- <Required verification evidence.>
## Context and exact contracts
<Reference existing requirements rather than copying a transcript. Include only
the relevant repository patterns, types, error cases, and interface
signatures. For a cheaper model: exact paths and signatures, a short example of
the expected shape, and the neighboring file to imitate.>
## Lessons that apply
<Rules copied from .crack/lessons.md, or "none".>
## Files and ownership
May change: <exactly this task's allowed_paths.>
Must not change: <other tasks' scopes and unrelated user work.>
Default exclusions: secrets and .env files, production config, undeclared
dependencies or lockfiles, CI, unrelated migrations, router or Codex
configuration, and .git internals.
## Implementation freedom
<Internal decisions left to the worker.>
## Verification
Working directory: <path>
Commands and expected results:
- `<actual command>` -> <expected result>
Required test cases: <happy, negative, and boundary cases.>
UI or runtime checks: <observable states, or why not applicable.>
If a check fails, reproduce it, diagnose the cause, and verify the in-scope fix.
## Stop rules
Continue the in-scope test/code/fix loop on your own. Report missing contracts,
unsafe conflicts, an unavailable environment, or repeated failures instead of
inventing requirements. No nested agents or CLIs, no permission changes, and no
commits, merges, or deploys.
## Escalation context
<Fallback rung only: the failed attempt's report and the review findings,
verbatim. Delete this section for a primary rung.>
## Required return
Use task-report.md: STATUS (ready_for_review, blocked, or failed), the actual
commands and exit codes, changed files including untracked files, acceptance
criterion -> implementation -> evidence locations, unresolved risks, and a resume
checkpoint. Keep full logs in task-owned files; surface failures and missing checks. Only the orchestrator accepts a task.
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 |