harness-engineer
claudeTurn any new or existing repo into a harness-ready codebase where coding agents can reliably ship features, pass tests, and maintain quality without human supervision
4 skills · 3 commands
Source-grounded scenario simulation, observable-output tracing, failure-mode synthesis, and fix blast-radius analysis. Mentally simulate workflows through the local code, trace what receivers actually observe, cluster failures by shared root cause and missing invariant, design durable fixes, then trace downstream repercussions across producers, consumers, APIs, events, cache, UI, permissions, and tests before implementation.
Plugin gedankenexperiment from marketplace enkira-plugins in enkira-ai/claude-plugins. Commands read from the repo's manifest — 17 days ago.
gedankenexperimentUse when asked to mentally simulate workflows, imagine realistic or adversarial scenarios, harden a system, find edge cases, audit state transitions or externally observable outputs, review call/chat/order/scheduling/payment/onboarding flows, design durable fixes, trace downstream repercussions and
Turn any new or existing repo into a harness-ready codebase where coding agents can reliably ship features, pass tests, and maintain quality without human supervision
4 skills · 3 commands
List and browse all Claude Code chat sessions with last access date, session ID, project, and name
2 skills · 2 commands
Shepherd a whole GitHub Epic (issue with sub-issues) end-to-end. Phase 1 snapshots open sub-issue topology + blocked_by edges into a tier-ordered sequence JSON. Phase 2 (/loop /epic-shepherd:shepherd) is the autonomous tick loop — dispatches an implementer for each ready issue, polls for the opened PR, hands it off to pr-shepherd to drain reviews + merge, regenerates the snapshot so closed issues advance the frontier. Phase 3 (v0.3.0) autonomously closes out the epic when the sequence empties: composes a close-out report, opens a PR that squash-closes the epic via 'Closes #N', shepherds that PR via pr-shepherd, exits the loop. Hard stops (human CHANGES_REQUESTED, branch protection, conflict, ping-pong cap, implementer failure, sub-issue reopened mid-closeout) pause and surface.
1 skill · 2 commands
Overnight PR grind — creates PRs from ready issues, addresses reviewer feedback (humans + bots like Copilot/Claude/Gemini), stops at 'ready to merge' (never auto-merges). Works with GitHub and gitlab.com.
1 skill · 1 command
Non-interactive second-opinion reviews via the Codex CLI. Wrap `codex review` for uncommitted changes, branch diffs, or specific commits — use for RFCs, PR prep, or any work where an independent cross-model reviewer would catch blind spots Claude's own review misses.
1 skill · 1 command
Scheduled dev briefs by email/Slack, one profile per teammate: the day's meetings, PRs awaiting their review ranked by urgency, PR blockers and progress, where each person stopped (the owner's Claude sessions, or a teammate's GSD agent state), a daily progress and design digest of chosen repos, a tracked effort's epic progress, each person's chosen topics across the org, Slack/email that needs them, the latest meeting notes and recordings (filed one folder per meeting in Drive), and PRs merged/opened in watched repos
1 skill · 1 command
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
15 skills · hooks
Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.
1 MCP server
Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.
5 commands
Automates browser interactions for web testing, form filling, screenshots, and data extraction
12 skills
Use Alibaba Open Code Review from local Codex.
2 skills
Claude Code skill for creating and presenting Slidev slidedecks
1 skill
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.