AI & LLM Engineering · By aiming-lab

Diagnoses where an agent failed across runs and turns the findings into new skills, system prompt patches and knowledge entries, using the A-Evolve loop.

aiming-lab/AutoResearchClaw

A skill your agent uses when implementing any endpoint, form handler, CLI tool, or function that accepts external input.

aiming-lab/MetaClaw

A skill your agent uses when writing shell scripts, Python automation, or any unattended batch job.

aiming-lab/MetaClaw

A skill your agent uses when writing messages in async channels (Slack, GitHub issues, email threads) where the reader may not have context and cannot ask follow-up questions immediately.

aiming-lab/MetaClaw

A skill your agent uses when writing any explanation, documentation, or response that will be read by someone else.

aiming-lab/MetaClaw

A skill your agent uses when exploring an unfamiliar codebase, tracing code paths, or answering questions about how the system works.

aiming-lab/MetaClaw

Best practices for image classification tasks. An agent skill from aiming-lab/AutoResearchClaw.

aiming-lab/AutoResearchClaw

Best practices for LLM alignment techniques including RLHF, DPO, and instruction tuning.

aiming-lab/AutoResearchClaw

Best practices for language model pretraining and fine-tuning.

aiming-lab/AutoResearchClaw

Best practices for reinforcement learning policy optimization.

aiming-lab/AutoResearchClaw

A skill your agent uses when diagnosing a bug, unexpected behavior, test failure, or any situation where code does not behave as expected.

aiming-lab/MetaClaw

Best practices for object detection tasks. An agent skill from aiming-lab/AutoResearchClaw.

aiming-lab/AutoResearchClaw

Multi-GPU and distributed training patterns with PyTorch DDP.

aiming-lab/AutoResearchClaw

Use FP16/BF16 mixed precision to accelerate training and reduce memory.

aiming-lab/AutoResearchClaw

Best practices for building robust PyTorch training loops. An agent skill from aiming-lab/AutoResearchClaw.

aiming-lab/AutoResearchClaw

Use this skill before executing a sequence of 3 or more steps, especially when steps are irreversible or depend on each other.

aiming-lab/MetaClaw

A skill your agent uses when presenting information from external sources, citing research, or answering factual questions.

aiming-lab/MetaClaw

A skill your agent uses when implementing a new feature or fixing a bug.

aiming-lab/MetaClaw

A skill your agent uses when deciding which tools to call in an agentic workflow.

aiming-lab/MetaClaw

Use this skill before taking any action that is hard to reverse — deleting files, overwriting data, sending messages, pushing to remote, modifying production systems.

aiming-lab/MetaClaw

A skill your agent uses when creating charts, plots, or dashboards.

aiming-lab/MetaClaw

Common mistake — stating specific facts (API endpoints, library versions, config options, function signatures) with false confidence when uncertain.

aiming-lab/MetaClaw

Common mistake — doing unrequested work (refactoring, adding extra features, cleaning up style) when the user asked for a specific, targeted change.

aiming-lab/MetaClaw

A skill your agent uses when the user's request is ambiguous, under-specified, or could be interpreted in multiple ways.

aiming-lab/MetaClaw

A skill your agent uses when a tool call, command, or API request fails.

aiming-lab/MetaClaw

A skill your agent uses when the user has a list of tasks and needs help deciding what to do first.

aiming-lab/MetaClaw

A skill your agent uses when summarizing progress on an ongoing project or multi-step task.

aiming-lab/MetaClaw

A skill your agent uses for any problem that involves multiple steps, tradeoffs, or non-trivial logic.

aiming-lab/MetaClaw

A skill your agent uses when a user presents a large, vague goal.

aiming-lab/MetaClaw

A skill your agent uses when you are not sure about a fact, have outdated knowledge, or the question is contested.

aiming-lab/MetaClaw

Common mistake — proceeding with assumptions about ambiguous requirements instead of asking a clarifying question first.

aiming-lab/MetaClaw

Common mistake — retrying the same failing command or API call without understanding why it failed.

aiming-lab/MetaClaw