Best context engineering skills, page 2

Keeps task_plan.md, findings.md and progress.md on disk as the agent's working memory for multi-step work, wired into Kiro steering, with no hooks.

OthmanAdi/planning-with-files

Reference for MoAI-ADK's core development principles: TRUST 5 quality gates, SPEC-first domain-driven workflow, agent delegation and token budgeting.

modu-ai/moai-adk

Applies a 12-stage verified workflow, from research to deploy, to non-trivial coding tasks, scaled to lightweight, standard or full mode by task size.

artemiimillier/bulletproof

Checks and tunes token-saver output compression: reads stats, explains why a command was or was not compressed, and edits config files or environment variables.

ppgranger/token-saver

A skill your agent uses when starting Claude Code projects, writing CLAUDE.md/spec.md, dispatching subagents, or requesting Agent Teams parallel development.

sangrokjung/claude-forge

Saves the current session as a searchable log with a curated summary and the raw transcript, so a later session can resume from it.

EliaAlberti/cpr-compress-preserve-resume

Acts on a Caveman learn report: reviews ranked token sinks, applies cost-lowering edits one at a time with your consent, and reports what each fix returned.

JuliusBrussee/caveman

Configure any third-party model as a custom model in Factory Droid (~/.factory/settings.json): resolve specs (context window, max output tokens, reasoning efforts), write a safe config entry, set…

nikships/droidproxy

Open the Cross-Code Organizer (CCO) dashboard — view and manage all memories, skills, MCP servers, hooks, and configs across scopes

mcpware/cross-code-organizer

Starts a background daemon that watches a Claude Code session's size and prunes it before auto-compaction can discard context or agent-team state.

Ruya-AI/cozempic

CRITICAL - Guide for using Claudish CLI ONLY through sub-agents to run Claude Code with any AI model (OpenRouter, Gemini, OpenAI, local models).

MadAppGang/claudish

Upgrade any Apple Foundation Models Swift app or package from OS 26-era APIs to OS 27 and Xcode 27 APIs.

rryam/FoundationModelsKit

Read or change compact-adviser settings for Codex - the mode (hint or off), the minimum context before a checkpoint is judged, the context budget, TypeSafe request logging, and the TypeSafe API key.

kunchenguid/compact-adviser

Audits a project's context routing, skill discoverability and skill wiring by running a chain of validator scripts and fixing whatever they report.

withkynam/vibecode-pro-max-kit

Uses Jev to mark each transcript turn keep, summarize or drop when cutting a conversation to a fixed size, with measured results on handoff quality.

kerpopule/hermes-jev-skills

Re-grounds a long-running agent in the current product outcome by running a path-based trace to the root of its topology and work trees.

mvschwarz/openrig

Reorganizes an overgrown MEMORY.md into a short pointer index plus separate topic files, and fixes or deletes outdated memories instead of archiving them.

simbajigege/book2skills

Determine which test suites are affected by recent code changes and run only the relevant ones.

GlitterKill/sdl-mcp

Search context-mode's persistent FTS5 knowledge base for previously indexed local project content, documentation, or session memory. Trigger…

mksglu/context-mode

Analyze Claude Code session bloat — shows token count, context usage %, and bloat breakdown.

Ruya-AI/cozempic

Sweeps a conversation before compaction and saves durable facts, decisions and session records into the Compartment encrypted memory vault as short one-claim memories.

MaxFreedomPollard/Compartment

Harness patterns for coding agents — memory, permissions, context engineering, delegation, skills, hooks, bootstrap.

keli-wen/agentic-harness-patterns-skill

This skill should be used when agent work needs file-backed context: durable scratchpads, tool-output offloading, just-in-time discovery, cross-agent handoff files, filesystem memory, or cleanup…

guanyang/open-agent-hub

Persistent, token-efficient project memory. An agent skill from shobcoder/shob.

shobcoder/shob

Surgically select, compress, and recover codebase context using Entroly's MCP tools.

juyterman1000/entroly

Estimates the per-turn token cost of a project's .claude folder and CLAUDE.md, split into always-loaded, path-scoped and invoked-only files, and flags what runs over budget.

poshan0126/dotclaude

Keeps a long Claude Code task on-track — breaks out of looping/circular thinking, watches the context budget, bounds internal reasoning, and triggers a clean handoff before the window fills.

thananon/9arm-skills

Audits a project's Claude Code setup for common token-cost leaks and returns a prioritized fix list, using only static inspection of files and settings.

mergisi/awesome-openclaw-agents

Coder-agent working-file budget discipline: keep the editable working set (files you /add into writable context) under ~25k tokens, separate "read" from "edit", delegate breadth to a read-only…

agentsope/SkillAlchemy

Build batteries-included agents with planning, context management, subagent delegation, and sandboxed execution.

langchain-ai/docs

Prunes a bloated Claude Code session with the cozempic CLI using a gentle, standard or aggressive prescription, after a dry run and your confirmation.

Ruya-AI/cozempic

Sweep the current session for uncaptured durable knowledge, file it to disk, persist the open work records this session knows are unfiled or now wrong, and curate the home's tiered, decaying startup…

kunchenguid/firstmate

Comprehensive instructions for executing tasks using the gws (Google Workspace) CLI or similar agent-first command-line tools.

google-labs-code/jules-sdk

Arabic edition of a file-based planning skill that keeps task_plan.md, findings.md and progress.md on disk so multi-step agent work survives lost context.

OthmanAdi/planning-with-files

Analyze the token length of an OT-Agent conversation-format (ShareGPT-style) dataset — the per-trace distribution (median/p90/max) and/or counts under a token threshold + a metadata predicate (e.g.

open-thoughts/OpenThoughts-Agent

Add a new LLM model to the preset catalog, or a whole new LLM provider — with the mandatory pre-flight verification of model id, pricing, and context window against the provider's official docs.

bagofwords1/bagofwords

Use Entroly for content-blind agent-history audits, explicit baseline/optimized trials, recoverable command or browser evidence, response contracts, and token-efficiency claim verification.

juyterman1000/entroly

Rewrites a memory file such as CLAUDE.md or a todo list in terse caveman-style text to cut input tokens, saving a readable backup outside the project tree.

JuliusBrussee/caveman-code

Clones only the later half of the current Claude Code session into a fresh one, cutting token usage while keeping the recent work intact.

ykdojo/claude-code-tips

Shared loop-engineering reference for COG skills - the agent loop, deterministic verifiers, termination conditions, in-loop context management, and named patterns.

huytieu/COG-second-brain

Routes a request to one of eight specialist agents covering hooks, skills, subagents, MCP integration and context engineering for Claude Code.

ohmyjahh/xquads-squads

UTILITY SKILL — Two-mode context-window management. An agent skill from jonathan-vella/apex.

jonathan-vella/apex

A developer guide to adding compact memory to an agent: when to trigger compaction, how to fork a compactor sub-agent, what the summary holds, and how to restore it.

simbajigege/book2skills

Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques.

Orchestra-Research/AI-Research-SKILLs

Answers questions about a named symbol, such as its callers, what it calls, the path between two symbols or the downstream impact of changing it, using the ripwire CLI.

redhat-et/ripwire

Add, change or verify a Nexus LLM model definition — the registry entry, the provider default, and proof the model id actually works against the live endpoint.

ProfSynapse/nexus

Builds a structural map of a large or unfamiliar codebase by dispatching sub-agents to summarize regions, keeping the main context small.

yologdev/yoyo-evolve

Rules for sending data-heavy work through context-mode's sandboxed MCP tools so raw output, web pages and large files stay out of the context window.

mksglu/context-mode