Agent Skills Hub

youtube-transcript-analyzer

Download and analyze YouTube video transcripts to extract technical insights, summarize complex tutorials, and relate video content to your codebase.

mess-mcp

An MCP server enabling Claude to dispatch and manage physical-world tasks using the MESS (Meatspace Execution and Submission System) protocol.

overview

Foundational mental model and operational rules for using TraceMem to ensure secure, auditable, and compliant AI agent execution.

langchain-architecture

Architect production-grade LLM applications using LangChain 1.x and LangGraph. Implement stateful AI agents, multi-step workflows, and custom memory systems for complex conversational and automation tasks.

planning-with-files

Implements Manus-style persistent markdown planning for complex workflows, project tracking, and research management to optimize agent attention and memory.

debug-distributed

Debugging guide for AReaL distributed training issues, including hangs, NCCL errors, OOM, and numerical consistency in FSDP2/TP/CP/EP.

libfuzzer

Coverage-guided fuzzer for C/C++ projects integrated with the LLVM toolchain.

serena

Symbol-level code understanding and navigation agent toolkit using LSP for precise code analysis, reference tracking, and surgical refactoring across 30+ programming languages.

owasp-mobile-security-checker

Automated security auditing for Flutter applications based on OWASP Mobile Top 10 (2024). Perform vulnerability scans for hardcoded secrets, insecure storage, dependency risks, and network configuration issues.

find-bugs

Perform automated security audits, bug detection, and code quality assessments on local branch diffs using a structured, checklist-driven verification process.

log-focus-debug

Focus debug skill for DashPlayer: isolates log chains, injects temporary focus markers ([FOCUS:token]), and ensures clean removal of debug artifacts after task completion.

context-degradation

Diagnose, isolate, and mitigate LLM context failures like lost-in-middle, poisoning, distraction, and context clash to improve agent reliability.