200 skills · sorted by relevance
Estimate remaining workflow costs from the average spend of completed steps. Show actual spending and projected totals in the status line, warn about configured budget overruns, and suggest a budget profile when estimated costs are high.
Reviews code for common security weaknesses, authentication problems, exposed credentials, and unsafe configuration. It adjusts review depth to the changes and offers an adversarial workflow, where defensive and attacking perspectives guide fixes and subsequent verification.
Override automatic work-context detection when research or writing is treated as software development. Keep workflows focused on market analysis, strategy, and documents throughout a session, or switch back to development mode or automatic detection.
Turn long videos into AI-curated short clips with OpusClip: create clip projects, check status, list clips with virality scores. Trigger phrases: opusclip, video clips, short clips, viral clips, clip a video.
The English summary is being prepared. Showing the original description.
Run fable-mode execution discipline on Claude Fable 5.1 — the top of the escalation ladder and the strongest staged run available. Routes the task to the @fable-orchestrator agent with its model overridden to Fable (`model: "fable"`), which stages the work, delegates artifact production to @fable-worker-sonnet / @fable-worker-haiku, and cold-checks deliverables with @fable-verifier. Trigger when the user explicitly asks for thorough/systematic/"deep work" handling on the strongest model ("fable on fable", "stage this on fable", "deep work mode, fable", "run this on 5.1"), or when a fable-opus run reports the task exceeds Opus's ceiling. Benchmarked 2026-09-02: zero correctness delta on Fable 5.1 at ~1.1× tokens, ~2× wall — a provenance runner, not an accuracy runner. Do NOT use for ordinary single-pass tasks — prefer fable-opus, fable-sonnet, or fable-haiku when the task doesn't need peak reasoning. Fable is plan-included; bulk work still routes down-tier.
The English summary is being prepared. Showing the original description.
Choose an agent and model tier using the task description and learned patterns. The skill can explain its recommendation, use compiler-based edits for supported simple changes, and record task outcomes to inform future routing decisions.
Run a prepared set of test cases to check agent-booster success rates and routing decisions. Optionally compare results with Gemini or Claude models, then save detailed measurements and summaries as JSON records for later verification.
Organize Happy's GitHub issues, find duplicates, and check whether reported bugs are fixed on main, the primary development branch. Prepare replies and proposed actions for maintainer review, with explicit approval required before changes affecting contributors.
Interpret the saved results of a completed AI council review. Count verdicts, check whether enough participants responded and whether models are independent, surface substantive dissent, and explain what action the outcome supports.
Research technical, market, or academic topics through multiple available AI providers working in parallel. Combine their findings into a report that preserves disagreements, identifies gaps, attributes sources, and recommends next steps with priorities.
Turn Markdown research and strategy outputs into office documents for sharing or presentation. Choose formats such as DOCX, PPTX, or PDF, organize sections for the audience, and apply consistent headings, tables, and slide formatting through document-skills.
Research technical or business questions through multiple available AI providers, each assigned a different perspective. Combine usable results into a report covering findings, disagreements, evidence gaps, priorities, and recommended next steps.
Keep a persistent record of project bugs and blockers in a local Markdown file. List issues by severity and category, inspect individual records, and save resolution notes so problems and their fixes remain available across sessions.
Review system architecture using implementation, callers, and tests as evidence. Compare interface designs, clarify how components should interact, and receive a recommendation with migration steps and test implications before deciding whether to refactor.
Resume a previous workflow after a new conversation or context reset. Read saved project state and handoff notes, summarize recent progress and blockers, and identify the appropriate next action for planning, implementation, or delivery.
Clarify a feature’s requirements with multiple AI providers, combining technical constraints with user and business needs. Produce a scoped definition covering priorities, edge cases, exclusions, and a requirements checklist before implementation begins.
Run a build-and-evaluation pipeline from a natural-language specification. Generate scenarios, implement the project, test scenarios withheld during implementation, and report scores against a target with failures and suggested next steps.
Breaks large changes, migrations, or multiple fixes into clearly scoped work packages. Independent workers use separate Git working directories, run checks, and report results, while dependency validation determines which packages can run together.
Save design documents with branch, author, and revision history so later sessions can recover earlier decisions. Keep previous versions intact, search related designs, and carry constraints and recommendations into implementation or review.
Review code or documents with available AI providers before delivery. Combine findings into a scored validation report that checks prior requirements and decisions, separates critical issues from warnings, and recommends next steps.
Review code with multiple AI perspectives to identify quality, security, performance, and design issues. The workflow also examines test evidence and unfinished implementation, then combines findings into a review that can be shared on a pull request.
Coordinates multiple AI providers to research, define, implement, and validate development tasks. It supports parallel task decomposition, comparison of proposed solutions, cross-model critique, and quality checks that feed into combined reports and final deliverables.
Create a structured specification in plain language from a feature description and multi-AI research. Define actors, behaviors, constraints, dependencies, and acceptance conditions, while checking completeness and keeping unresolved user decisions visible for later review.
Review code independently with five AI models, combine their reports, and filter out duplicate or unsupported findings. Present concrete bugs and risks for approval, then fix only the issues the user accepts.