| --- | |
| name: tla-verification | |
| description: TLA+/PlusCal formal verification for concurrent protocol design. Use when modeling async systems, checking race conditions, or verifying state machine correctness before implementation. | |
| --- | |
| # TLA+ Protocol Verification — Pilot's Checklist | |
| You are a formal verification specialist. You help design, write, debug, and run PlusCal/TLA+ specifications for concurrent systems, with a focus on asyncio-style cooperative concurrency patterns. You know the sharp edges of PlusCal, TLC model checking, and how formal models map to real implementation code. | |
| ## When to Use This Skill |
| ok I'd liek to undertsand the memory thoroughly in hermes agent. Start with docs and undertstand hermes agent architecture amnd then dive into memory first in docs and then in code. | |
| I'd like to deploy memeoy in my applicaytion and i tyhink hermes agent ha really interesting approach to memory | |
| ⏺ Let me explore the codebase systematically - starting with docs, then architecture, then memory implementation. | |
| Searched for 2 patterns, listed 1 directory (ctrl+o to expand) | |
| ⏺ Good - I can see there's rich documentation. Let me read the architecture doc and memory-related docs in parallel. | |
| Read 4 files (ctrl+o to expand) |
| #!/usr/bin/env bash | |
| # ────────────────────────────────────────────────────────────────────── | |
| # litellm supply-chain attack auditor (CVE pending — 2026-03-24) | |
| # | |
| # Checks for indicators of compromise from litellm 1.82.7 / 1.82.8 | |
| # which were published to PyPI with a malicious .pth payload that | |
| # steals credentials, exfiltrates to models.litellm.cloud, and | |
| # attempts lateral movement via Kubernetes. | |
| # | |
| # Safe versions: anything < 1.82.7 |
Fine-Tuning Sparse Encoders for Neural Sparse Retrieval: Complete Reproducible Guide (SPLADE, OpenSearch Neural Sparse, CSR)
Last updated: 2026-02 Scope: Covers SPLADE / SPLADE++ / SPLADE-v3, OpenSearch Neural Sparse v1–v3 (including inference-free), and CSR (Contrastive Sparse Representation). Includes full training recipes, loss functions, architecture internals, and practical tips.
| #!/usr/bin/env python3 | |
| """Extract Claude Code session JSONL to lean readable markdown. | |
| Usage: python3 claude_session_extract.py <session.jsonl> [output.md] | |
| If output.md is omitted, writes to <session_id>.md in current directory. | |
| """ | |
| import json | |
| import sys | |
| from pathlib import Path |
| #!/usr/bin/env python3 | |
| """ | |
| RLM v2 (Recursive Language Model) Module for Smolagents | |
| Implements DSPy RLM-style strategies for handling large contexts: | |
| - Peeking: Look at data structure before processing | |
| - Grepping: Use string/regex matching before LLM calls | |
| - Partition + Map: Chunk data and process with batched sub-LLM calls | |
| - Summarization: Hierarchical compression for understanding the whole | |
| - Hybrid: Combine strategies as needed |
| #!/bin/bash | |
| # Crawl and download Claude platform docs as markdown | |
| set -e | |
| BASE_URL="https://platform.claude.com/docs/en" | |
| SITEMAP_URL="https://platform.claude.com/sitemap.xml" | |
| WORK_DIR="/tmp/claude-docs" | |
| OUT_DIR="$WORK_DIR/docs" | |
| URLS_FILE="$WORK_DIR/urls.txt" |
| """ | |
| SDK Patch with Subagent Support - Adds .raw field to SDK message types with maximum fidelity. | |
| Provides drop-in replacements: | |
| - ClaudeSDKClientWithRaw: replaces ClaudeSDKClient | |
| - query_with_raw: replaces query | |
| Data sources for .raw: | |
| - user/assistant: JSONL file (has parentUuid, timestamp, isSidechain, etc.) | |
| - result: Raw CLI output (has modelUsage, errors, permission_denials) |
Authorship Note: This document was compiled during an interactive exploration session simulating a "Feynman Lab" environment. It deconstructs the
Luxicalproject to explain how modern engineering (Rust, Numba, Distillation) allows simple arithmetic to achieve state-of-the-art results.