Намерете инструкция, прочетете оригиналния файл и подгответе команда за своя инструмент.
31 301 skills в каталога
Как подбирам и измервамДанните имат нужда от нова проверка.Последна успешна проверка: 23.09.2026 г.
Проверката на източника не означава, че всяка инструкция е изпробвана от Level8. Каталогът не обхваща непременно всички публични хранилища.
Търсене в целия индекс. Метаданните показват посоченото от източника; липсващите стойности остават означени.
Датата в GitHub е отделна от първата поява в моя индекс. Филтърът по дата изключва записите без посочено обновяване.
Страница 196: 24 записа в 10 групи по репозиторий. Всички записи от страницата са показани. Номерът пази позицията от избраната подредба.
12 skills от този източник на тази страница
Validate and reconcile backend-neutral agent lifecycle state, explicit capability support, and repo-relative mutation scopes. Use before routing, resuming, messaging, or running concurrent mutating agents.
Memory as a learnable skill — store, recall, compress, and consolidate agent memories. Inspired by Stanford AutoMem.
Auto-decide whether a task runs on a LOCAL model or a CLOUD model (DeepSeek, GLM, Nemotron, Grok, Gemma, …) from an automatic effort estimate, and run multi-model fusion (cascade, draft→refine, vote). Use when the user wants automatic local-vs-cloud routing, to add cloud LLM providers, to make local and cloud models collaborate, or mentions effort-based routing, model fusion/ensemble, OpenRouter, DeepSeek, GLM, Nemotron, or Grok.
Run inference with tiny feedforward classifiers and maintain their learning loop — predict with one model, classify across all models, list available models, plus active learning, temperature calibration, expected calibration error and dataset auditing; executes through a compiled Rust binary with a numpy fallback. Use when a routing or effort decision should come from a small trained classifier rather than a prompt. Not for the tier decision itself — use nn-router; not for a large language mode
Wrap an installed coding agent so its traffic goes through the local compression proxy, and unwrap it again — wrap, unwrap, list, status, with the proxy supervised on its port and the agents discovered on the machine. Use when an existing agent binary should benefit from compression without editing the agent. Not for deciding what to compress — use token-shaper (policy) or universal-compressor (content); not for installing cross-agent MCP plugins — use plugins; not for measuring the resulting sa
Cache a project's scan result so the first agent scans and the following agents read the cache instead — get_or_scan, a separate audit report slot, a .botte-cache store and a 24 hour TTL. Use when several agents work on the same project in sequence and the scan is the repeated cost. Not for caching a model response by question — use response-cache; not for caching an agent run whose output is predictable from its input — use agent-cache; not for skipping unchanged files during an audit — use cod
Run the canonical, already-optimal project checkup in one command — policy presence, directives health, per-component metrics, infra tips, duplication, recurring Markdown review, Wiki/second-brain integrity, and drift detection — so you never have to hand-write a good checkup prompt. Use when the user says 'do a complete checkup', 'botte doctor', after a component update, when onboarding to a project, or when multiple agents/devs may have caused drift.
Treat the homelab/micro-cluster as one schedulable resource — discover every reachable machine and spread cheap work across them (least-recently-used first) so idle boxes get the next task, recovering wasted local capacity instead of paying the cloud. Use to see cluster status, route a task to the idlest machine, or hand a task to a trusted agent on another machine. Also use when the user mentions homelab, cluster, idle machines, load balancing, or distributing work across machines.
Pick the optimal set of skills/docs to load for a task under a token budget — an exact 0/1 knapsack (maximize relevance while summed token cost stays under budget), not an LLM 'decide what's relevant' call. Deterministic, stdlib, 0 cloud tokens. Cuts the always-on context cost (loading the whole skill catalog every turn). Use when you want to load only the most relevant context within a budget, or to decide which skills/docs an agent should read for a task. Not for measuring how much the prefix
Compress a large log file before an agent reads it — collapse numeric, timestamp and UUID noise into patterns, count repetitions, keep the top patterns and a sample of unique lines, with a reported 80 percent reduction on log-analysis tasks. CONSOLIDATION PENDING: this overlaps the log strategy of universal-compressor (content_type log, 80 to 98 percent), which is the supported path for new work; this module is kept for existing callers only. Use when a log has to be reduced to its patterns and
Measure a project's always-on prefix (agent directives + core rules + MCP tool schemas + skill catalogue) in tokens and as a % of small local-model windows (64k/128k/256k), with a concrete reduction plan (lazy tool loading, on-demand skill search). Use to see how much of a modest machine's usable context is spent before any real work, and how to shrink it so weaker machines can run local LLMs usably. Not for choosing what to load for one task — use context-budget; not for deleting sections that
Estimate what a task or a fix will cost — tokens, model/tier, money ($), and wall-time — using the tiered cost model. Use whenever the user wants to know the cost of a correction, audit, or task before running it, or to compare local vs cloud cost.
1 skill от този източник на тази страница
Profile CSV files locally: missing values, duplicate rows, columns, encodings and data quality. CSV 결측 중복 데이터 품질 검사.
1 skill от този източник на тази страница
Discover and set up released Legends modules for rank grids, search research, GitHub improvements, audio generation, OBS recording, cursor overlays, and voice typing, and source-cited Obsidian vault memory. Use for cto-legends, ecosystem updates, or matching these goals to installable tools.
2 skills от този източник на тази страница
Use when building a chart. Pick the form from the question being asked, then refuse the distortions: truncated bars, dual axes, mean-only summaries, and colour used as decoration.
Use when choosing which AI model to use for a task. Match model class to the axis the task actually stresses, then measure on your own prompts.
1 skill от този източник на тази страница
Mine a Discovery App workspace's .discovery/ exhaust for token usage across engines (Clio / Science Engine, copilot-cli / Mission Control / Generic Copilot), models, and interactive chat. Produces a console read-out by default; can also emit a markdown report on request. Cost / billing-multiplier da
1 skill от този източник на тази страница
Socratic coaching grounded in the user's journal, informed by attachment theory, schema therapy, polyvagal theory, and Big Five personality. Philosophical, clinical, confrontational when needed, always compassionate. Reads journal-coach.md for user metadata.
1 skill от този източник на тази страница
Use when the user wants to free disk space on macOS, find leftover files from uninstalled apps, audit what is installed, or figure out why their Mac is slow or full. Scans every install surface (Homebrew, npm, .pkg installers, curl installers, language toolchains, app support data), proves what is actually used before suggesting anything, then removes only what the user confirms.
2 skills от този източник на тази страница
Calibrates Marlin 2.x 3D printer firmware bed leveling: Unified Bed Leveling (UBL), Bilinear ABL, M420 S1 post-homing, Z-probe offsets, G26 mesh prints, and EEPROM slots. Trigger phrases: marlin bed leveling, ubl calibration, m420 s1, z probe offset, g29 bed level, first layer adhesion.
Calibrates Marlin 2.x 3D printer firmware bed leveling: Unified Bed Leveling (UBL), Bilinear ABL, M420 S1 post-homing, Z-probe offsets, G26 mesh prints, and EEPROM slots. Trigger phrases: marlin bed leveling, ubl calibration, m420 s1, z probe offset, g29 bed level, first layer adhesion.
2 skills от този източник на тази страница
Validates 1EdTech / IMS Global OneRoster v1.1 and v1.2 CSV roster sets: manifest integrity, bulk vs delta strictness, foreign key references, and encoding sanitization. Trigger phrases: oneroster csv validator, validate oneroster zip, clever roster error, oneroster sourcedid, bulk vs delta oneroster.
Validates 1EdTech / IMS Global OneRoster v1.1 and v1.2 CSV roster sets: manifest integrity, bulk vs delta strictness, foreign key references, and encoding sanitization. Trigger phrases: oneroster csv validator, validate oneroster zip, clever roster error, oneroster sourcedid, bulk vs delta oneroster.
1 skill от този източник на тази страница
The opponent role — a red team for any decision. Called at a fork: it tries to break the decision against data, checks the premises, the economics and the second-order consequences, and produces a review the author answers line by line. Read-only, changes nothing. Works in any session with any author, human or another agent, and needs no project context. Best in a separate session with none of the author's context, or on a model from a different vendor.
Пазя съществуващия каталог и добавям намерени SKILL.md файлове след прочит на метаданните. Наличният индекс остава използваем, когато проверката е частична.
Платформата и нивото описват записаното в източника. Те не са гаранция за съвместимост. Български текст се показва само когато в каталога има отделен превод; иначе виждате оригинала с означен език.
Общите звезди показват натрупана популярност. За „Набират интерес“ се сравнява измереният темп при достатъчен интервал. ≥ означава непълна извадка. Под един час не изчислявам дневен темп; методът и изминалият период се показват до записа.
Проверете файла, лиценза и нужния достъп преди употреба. Нито каталогът, нито четенето на SKILL.md изпълняват външни инструкции.
Темата Agent Skills в GitHub