Menu
Skills dans ce dépôt
SkillsMP a collecté 447 skills depuis jleechanorg/claude-commands. Ouvrez un skill pour examiner sa source et ses détails.
jleechanorg/claude-commandsAffichage de 40 skills collectés sur 447.
Triple-blind AI testing protocol for scientifically rigorous experiments that test AI behavior changes (e.g. CLAUDE.md variants, prompt strategies) while eliminating coordinator and evaluator bias. Use when the user wants to run /experiment, design a…
Langue du texte source : anglais
Read-only fleet-size consumer for the 6 mac + 16 linux ezgha runner invariant. Restart-only remediation is UNAVAILABLE (fail-closed) pending jleechanorg/ez-gh-actions PR 67 (dual-Lima convergence) and PR 70/issue 60 (recovery controller) merging AND being…
Langue du texte source : anglais
Analyze conversation and git history to find gaps where cold reviews (codex, Bugbot, CodeRabbit, /reviewdeep) caught issues that dark-factory in-pipeline reviewer nodes missed. Fans out subagents, opens PRs, drives each through /green, and merges (with…
Langue du texte source : anglais
Display the Dark Factory pipeline node graphs, gates, node types, edge conditions, and handler mappings.
Langue du texte source : anglais
Use when a persistent repository, tool, configuration, automation, launcher, wrapper, or installed CLI fix could remain only in a working tree, topic branch, or one machine's live state.
Langue du texte source : anglais
Analyze and fix PR blockers (CI failures, merge conflicts, bot/review feedback) to get a PR into a mergeable state — never merges it. GitHub status is the sole source of truth over local assumptions. Slash command /fixpr. Also embodied as the copilot-fixpr…
Langue du texte source : anglais
Evidence-based E2E test generator for your-project.com (Real Mode Only). Use for the /generatetest slash command — generates self-contained MCPTestBase test files under testing_mcp/ with built-in evidence bundle generation (git provenance, checksums,…
Langue du texte source : anglais
Goal-driven harness loop — keep working on a goal until /es, /er, /code_standards, and Codex plugin review all pass. All gates use adversarial subagents. Alias: /h
Langue du texte source : anglais
Use when auditing or optimizing CLAUDE.md, AGENTS.md, GEMINI.md, slash commands, skills, or other coding-agent harness surfaces, or when recurring agent failures suggest instruction, policy, memory, test, or automation gaps.
Langue du texte source : anglais
Harness-fix meta-skill (slash: /meta) — analyze an agent behavior failure (autonomy violation, refusal, premature stopping) and run /harness to fix the agent, NOT the underlying task. Input: Slack thread URL, pasted conversation, or freeform description.
Langue du texte source : anglais
Rotate or repair Hermes Slack credentials, provision the complete Hermes Slack app permission baseline from the beginning, update macOS shell exports safely, and verify the live gateway path. Use for Hermes Slack token rotation, Slack app reinstall, Socket…
Langue du texte source : anglais
Alert when the Mac runner host's free disk drops below 50GB, auto-clean safe targets (evidence bundles, scratchpads, merged-PR worktrees) below 20GB. Closes the gap where mac-runner-di[REDACTED_OPENAI_KEY] (retired) and ezgha's own docker-daemon-view disk…
Langue du texte source : anglais
Harden a stated goal into ironclad exit criteria (stronger than asked, binary, executable, externally anchored, anti-gaming, iterate-until), set them durably (bead/goal/STATE/memory), then execute toward them. Use when the user invokes /ironclad, asks for…
Langue du texte source : anglais
Terminate macOS SecurityAgent and dismiss all stacked keychain modal prompts. Programmatically dismiss active macOS credential popups and debug headless/background keyring access issues (launchd, cron, isolated tmux test runners). Slash command:…
Langue du texte source : anglais
Capture durable learnings from failures, corrections, repeated mistakes, successful recovery patterns, or direct /learn requests. Use whenever the user invokes /learn, asks to save/remember/capture a lesson, or when a workflow failure should become reusable…
Langue du texte source : anglais
Validate that level-up flows are complete, auto-selected, editable, and evidenced with real LevelUpAgent output.
Langue du texte source : anglais
Trigger and monitor the level-up organic test (testing_mcp/core/test_level_up_organic.py) against the PR's deployed preview server.
Langue du texte source : anglais
Steering work on $USER's Ubuntu machine (jeff-ubuntu) via SSH. Use when asked to run commands, install software, check services, or manage files on the Ubuntu box.
Langue du texte source : anglais
Detect Gemini SDK / BYOK leaks when local AGY-default provider is expected.
Langue du texte source : anglais
Use when supervising the level-up ZFC migration loop in this repo, especially when the cleanup-first roadmap is drifting, AO workers need steering, or PR sequencing must stay aligned with the canonical roadmap.
Langue du texte source : anglais
Search across all memory systems — ~/roadmap, beads, claude memories, hermes sqlite, hermes briefings, hermes index, openclaw memories, wiki, history, and slack. Use whenever the user asks to search memories, find something in memories, or looks for anything…
Langue du texte source : anglais
Use for mobile browser investigations in this repo, especially Firebase/Google auth, iOS Safari or Chrome iOS, Incognito/Private Browsing, storage partitioning, redirect flows, and user-visible mobile-only failures.
Langue du texte source : anglais
Situational assessment, beads + roadmap sync after a work block; writes a self-contained nextsteps markdown doc (TOC, executive summary, full detail, bead links), updates learnings + README, Claude auto-memory, mem0, and beads. Prefers editing existing…
Langue du texte source : anglais
OpenClaw agent model configs — which work, which are broken/quota-limited, and how to switch
Langue du texte source : anglais
Autonomous convergence via orchestration — runs /converge repeatedly inside a tmux agent (managed by the existing /orch system) until success criteria are met, time limit reached, or max attempts reached, then executes a final /pushl → /reviewdeep → /copilot…
Langue du texte source : anglais
Run one benchmark script that compares pairv2, pair via Claude Teams, and pair via direct Python
Langue du texte source : anglais
Pairv2 implementation philosophy: LLM-decided outcomes, fail-soft file handling, and evidence-first promotion
Langue du texte source : anglais
Use when designing, debugging, reviewing, or scripting any work that has independent items (rows, files, tests, migrations, jobs, agent lanes, API sweeps, builds). Slash command `/parallel`. Enforces a single rule — the speed ceiling is the workload's real…
Langue du texte source : anglais
Run deterministic Playwright UI tests for local or test environments, including isolated browser profiles, traces, video, and multi-browser coverage. Use when the user asks for /playwright, end-to-end tests, reproducible UI testing, visual regression, or…
Langue du texte source : anglais
Lazy senior-dev mode — the seven-rung ladder that decides whether to write code at all, whether to reuse code that already exists, and whether to prefer stdlib/platform/installed deps over a new dependency. Use before writing any code, every PR diff, every…
Langue du texte source : anglais
Use this skill any time a .pptx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx file (even if the extracted content will be used…
Langue du texte source : anglais
Drive all open PRs toward /green (CI green + no merge conflicts) by fixing CI failures, and toward draft-phase quality readiness by resolving comments and running the smoke gate. CodeRabbit/Bugbot are optional advisory reviewers — surfaced for information,…
Langue du texte source : anglais
Use for production-code PRs ($PROJECT_ROOT/**, gates, ZFC). Requires full GitHub URL to governing roadmap/design doc and a `br` bead ID in the body.
Langue du texte source : anglais
Canonical /green definition — CI green + no merge conflicts, both verified at current PR HEAD SHA. Quality gates (evidence, review, comment-resolution) live in the draft-first-pr skill, not here.
Langue du texte source : anglais
This skill should be used when the user asks for a PR report, PR audit, PR delta analysis, /pr-report, or wants per-PR summary with delta files/lines. Generates a structured report for one or more open PRs covering purpose, files changed, +/− line counts,…
Langue du texte source : anglais
Interpret user-requested read-only work as protecting Git-tracked files while allowing commands and writes outside the Git index.
Langue du texte source : anglais
Canonical /repro workflow for copy+targeted bug repro, twin clones, evidence exports, same-symptom verdicts, and red/green provenance.
Langue du texte source : anglais
Run matched A/B reviewer calibration for /f: compare factory reviewer, raw codex exec, and delegated subagent reviews on the same frozen PR/work-item envelope.
Langue du texte source : anglais
Use this skill whenever a bug tempts you to add backend protection, fallback, clamp, sanitizer, retry, suppression, guardrail, or workaround logic. It enforces root-cause analysis first: inspect raw model prompts/responses and fix upstream prompt/schema/agent…
Langue du texte source : anglais
Operational health snapshot for the jleechanorg self-hosted runner fleet. Use when user says "check runners", "are runners up", "is jeff-ubuntu down", "diagnose runner health", or before/after runner operations. Produces a structured Markdown report at…
Langue du texte source : anglais