AI & ML

matthewye Skill

Autopilot task implementer. Reads AGENT-BRIEF, follows TDD discipline, auto-diagnoses errors.

matthewye Bundle

Autopilot issue resolution loop: scan → implement → review → retry. Dispatches implementer and reviewer subagents per ready-for-agent issue via the Agent tool, then runs global meta-review. Use when processing autopilot issues from any source.

l3rowniecakez Skill

[lab] v26.5.16 L-SKLL | Read and write to Oracle inbox — notes, tasks, messages, handoffs. Use when user says "inbox", "leave a note", "write to inbox", "check inbox", "what's pending", or wants to read/write messages for self or other agents. Do NOT trigger for session handoffs (use /forward), schedule (use /schedule), or agent messaging (use /talk-to).

l3rowniecakez Skill

[standard] v26.5.16 L-SKLL | Explore a codebase with parallel Haiku agents — clone, read, and document. Modes — --fast (1 agent), default (3), --deep (5). Use when user says "learn [repo]", "explore codebase", "study this repo", or shares a GitHub URL to study. Do NOT trigger for finding projects (use /trace), session mining (use /dig), or cloning for active development (use /incubate).

l3rowniecakez Bundle

[lab] v26.5.16 L-SKLL | Persistent agent mailbox — store findings, standing orders, and context for team agents across sessions. Use when user says "mailbox", "agent memory", "standing orders", "what did scout find", or wants to manage persistent agent knowledge.

l3rowniecakez Skill

[standard] v26.5.16 L-SKLL | Talk to another Oracle agent via contacts + threads. Use when user says "talk to", "message", "chat with", or wants to communicate with another agent (e.g. "talk to pulse", "message neo"). Do NOT trigger for OracleNet social feed (use /oraclenet), skill management (use /oracle), or family registry (use /oracle-family-scan).

l3rowniecakez Skill

[lab] v26.5.16 L-SKLL | Manage Oracle contacts — add, list, remove agents with their transport info (maw, inbox, thread). Use when user says "contacts", "add contact", "register agent", "who can I talk to", "list contacts". Do NOT trigger for sending messages (use /talk-to) or family registry (use /oracle-family-scan).

matthewye Bundle

Codex autopilot loop: scan -> implement -> review -> retry via spawn agent, then global meta-review.

matthewye Skill

Autopilot task implementer. Reads AGENT-BRIEF, follows TDD discipline, auto-diagnoses errors.

matthewye Skill

Autopilot issue resolution loop: scan → implement → review → retry. Dispatches implementer and reviewer subagents per ready-for-agent issue, then runs global meta-review. Use when processing autopilot issues from any source.

tommyducx Bundle

腾讯会议：会议管理与音视频协作助手。预约/创建/修改/取消会议、查询会议详情与会议号转换、查看参会成员/受邀人/等候室成员、查询用户会议列表（即将开始/进行中/已结束）、查询录制列表与下载地址、获取转写全文/段落/搜索、获取AI智能纪要（支持多语言翻译）、录制权限申请（预览+提交两步流程）、时间转换与版本检查、Agent意见箱反馈上报。当用户需要预约或管理腾讯会议、查看参会人员、查询会议录制或转写内容、获取智能纪要时使用；当用户访问录制相关内容（下载地址/转写/智能纪要）出现无权限错误时，自动发起录制权限申请流程；当Agent遇到以下任何情况时必须立即触发submit_feedback：工具缺失（功能不存在）、工具报错（返回错误码）、能力不足（参数/功能限制）、结果不符预期（数据异常）、改进建议（功能优化）；如果用户提供的是9位会议号而非meeting_id，优先使用get_meeting_by_code转换后再操作。不要在以下场景触发：日程管理（非腾讯会议日程）、即时通讯/聊天、腾讯文档操作、企业微信审批流程、电话/PSTN拨号、视频剪辑或视频编辑、其他视频会议平台（Zoom/Teams/飞书/钉钉）。

arozumenko Bundle

Use when a team wants to validate test-runner's bug-detection accuracy or test-author's authoring fidelity against their OWN app before trusting the manual-qa bundle on real work — "how good is this agent, really", "benchmark test-author/test-runner", "build a gold suite", "self-eval the manual QA team". Ports a held-out-answer-key eval methodology (deterministic Tier A scoring + a judged Tier B rubric) that keeps a self-authored eval honest, generalized to any app.

arozumenko Bundle

Use when profiling a mobile app, authoring or sizing mobile TCs, orchestrating a mobile suite run, executing a mobile TC, generating a manual guide, or producing a mobile run report. Mobile-app testing for all manual-qa agents — native iOS/Android via Appium MCP and the Mobitru device farm, and PWA/hybrid via Playwright mobile viewport.

arozumenko Bundle

Use when deriving a data-flow diagram and STRIDE threats from code, with a citation per element and mitigations recorded as claims a fresh reviewer can confirm or refute; provides the threat-model.json shape cite.mjs check validates.

arozumenko Bundle

Use when the user asks 'what did this cost', cost per session/role/test case, which role or sub-agent burned the most, tool-call/skill/time breakdowns, 'before vs after' cost comparisons, or wants to audit AI spend over time. Measures the token/cost/time efficiency of AI coding-agent work — per session, per role, per day, and per individual sub-agent — with every dollar metered by ccusage.

arozumenko Bundle

Use when a manual-QA agent does live browser testing through the Playwright MCP server — exploring/profiling a web app or executing a web test case against a running app, with no test code generated. Web/PWA/hybrid targets; used by app-profiler and test-runner. Mobile native → use mobile-testing instead.

swyxio Bundle

Reconstruct and visually analyze paired agent, game, or policy trajectories to determine whether changed actions produced their intended effects. Use for move-history audits, replay/trace analysis, policy regressions, behavior calibration, causal first-divergence studies, reply-survival analysis, agent personality validation, promotion decisions, and reports that must connect aggregate outcomes to exact state-action sequences.

swyxio Bundle

Build or improve progress, cancellation, result handoff, and reliability when creating or changing a user-facing slow asynchronous action, model call, media job, queue, or multi-step workflow. Applies to batch size, concurrency, and rerun changes affecting the visible wait; do not wait for a stuck-action complaint. Exclude ordinary fast requests and backend-only changes with no user-facing operation affected.

arozumenko Bundle

Use when running a security engagement end to end as the lead — init, review, register, model, cases, hand-off, report, fix, acceptances, sign-off; points at the other security-testing skills rather than duplicating their commands.

arozumenko Bundle

Use when asked to run a retrospective, mine past sessions, or improve the team from what already happened — turning prior Claude Code conversations and sub-agent sessions into proposed memory and workflow updates. Used by scout.

arozumenko Bundle

Use when a manual-QA agent does live browser testing through the Playwright MCP server — exploring/profiling a web app or executing a web test case against a running app, with no test code generated. Web/PWA/hybrid targets; used by app-profiler and test-runner. Mobile native → use mobile-testing instead.

swyxio Bundle

Build or debug stateful applications that specifically use Cloudflare's `agents` package, including Agent classes, routing, callable methods, synchronized state, scheduling, AIChatAgent, durable execution, or Agents SDK client hooks. Do not trigger for generic AI-agent design, the OpenAI Agents SDK, or ordinary Durable Object work that does not use the Cloudflare Agents SDK.

swyxio Skill

Add or review developer- and AI-facing conference data endpoints such as llms.txt, sessions.json, speakers.json, or MCP routes for AI Engineer conference pages. Implement only the requested surface unless the user asks for the full endpoint suite; never expose private speaker, CFP, or organizer data.

arozumenko Bundle

Use when turning a threat model's candidate cases or your own drafts into passive security test cases in manual-qa format, admitting them by lint against the allowed-operation grammar, and handing off a verified suite; provides cases.mjs.

arozumenko Bundle

Use when asked to run a retrospective, mine past sessions, or improve the team from what already happened — turning prior Claude Code conversations and sub-agent sessions into proposed memory and workflow updates. Used by scout.

swyxio Bundle

Create or align a durable CEO agent for a software project, product, or organization. Use when someone asks for a project CEO, product steward, general manager, autonomous owner, or open-ended agent that should proactively observe the product, set priorities, coordinate work, and report outcomes over time. Establishes evidence sources, authority boundaries, operating cadence, initiative and delegation rules, privacy protections, and a decision-oriented reporting contract.

swyxio Bundle

Critically audit or explicitly trim an existing agent skill to its behavioral core, including narrowing over-broad trigger metadata. Use when the user asks to cut, simplify, shorten, de-slop, or reduce the context cost or trigger aggressiveness of a SKILL.md or skill folder. Do not trigger merely because a skill is long or is being edited for another purpose.

swyxio Bundle

Design and implement AI copilots that analyze structured data and propose mutations for human approval (draft → apply), not direct writes. Covers prompting, multistep compound requests (remove + add in one turn), FIFO client queuing for long agent loops, validation, session memory, optimistic versioning UX, copilot panel/surface ergonomics (floating-dockable panel, summon-hotkey size state machine), LLM tracing/observability (OpenInference + OTLP session.id grouping, orphan-span fixes, human thumbs up/down feedback annotations), and test matrices. Use when building scheduling/admin chatbots, proposal workflows, human-in-the-loop agents, or reviewing aiebot-style features.

swyxio Bundle

Diagnose or improve reliability of a structured, multi-request, rate-limited, or cost-sensitive AI workflow. Use for malformed or truncated outputs, retry/rate-limit failures, unreliable fan-out, cache/resume bugs, or missing run telemetry. Do not use for ordinary prompt edits or simple one-shot model calls.

swyxio Bundle

Add or repair Cloudflare Turnstile on an existing web form by creating or reusing a widget, embedding it, wiring mandatory server-side Siteverify in the existing backend, and validating the result. Use when the user explicitly asks to add Turnstile, replace reCAPTCHA or hCaptcha with Turnstile, protect a form from bots, or fix a Turnstile integration. Do not trigger for unrelated Cloudflare or general form work.

00200200 Skill

Review an upstream documentation change against the skills, agent instructions, or runbooks that cite it. Identify supported updates, unaffected instructions, and unresolved version or evidence gaps. Use with a supplied source diff or Skill Watch result.

swyxio Bundle

Design or harden Slack bot transport and interaction architecture: signed Events API ingress, fast acknowledgement, idempotency, causal thread context, Block Kit actions, durable agent execution, multi-workspace operation, or Slack-specific delivery of typed follow-up controls and privacy-bounded execution receipts. Use for new Slack bots, broad Slack architecture audits, or changes to these Slack contracts. Do not use for deployment or visual testing, cosmetic copy/layout, general answer quality, channel-agnostic transcript or receipt schemas, core suggestion generation, or merely operating Slack.

swyxio Bundle

Create or align a durable, independent Cassandra agent for a software project, product, or organization. Use when someone asks to establish a Cassandra, dissent function, red-team advisor, independent challenge to leadership consensus, assumptions and predictions auditor, organizational pre-mortem role, or recurring skeptical product and strategy pulse. Establishes evidence access, independence from CEO or execution agents, a durable ledger, interruption threshold, read-only authority, reporting contract, dedicated thread, and optional approved automation.

swyxio Bundle

Build or refactor a long-running AI workflow that needs visible progress, durable partial results, resume after interruption, or atomic publication. Use for multi-stage extraction, indexing, batch analysis, evaluation, or report workflows. Do not use for a short synchronous model call or a UI-only streaming feature.

swyxio Bundle

Summarizes arbitrarily long text (1k-1M words) using recursive map-reduce with any LLM backend. Accepts raw text, markdown, transcripts, articles, codebases, or any plaintext input. Produces one or more output formats: executive summary, section headings with timestamps, YouTube description, Twitter/X posts, title options, thumbnail prompts, blog outlines, pull quotes, and more. Supports focus directives ("focus on the AI parts", "emphasize the business angle") to steer the summary. Pluggable backends: OpenRouter, Ollama, OpenAI, Anthropic, Gemini, or any OpenAI-compatible endpoint. Use this skill when someone says "summarize this", "give me a summary", "TL;DR", "make this shorter", "create a YouTube description", "write a tweet about this", "generate titles", "thumbnail ideas", or provides long text and wants any condensed output.

swyxio Bundle

Run Codex models and reusable Cursor, Antigravity, Muse, Deep Code, ZCode, Devin, or Mistral Vibe CLI adapters programmatically, with latency, error, usage, cost, and trace logging. Use for scripted summarization, structured extraction, classification, code generation, tool or installed-skill invocation, batch processing, or model comparisons when coding-agent CLI authentication and capabilities are required. Do not use when an ordinary interactive Codex turn is sufficient.