agent-mem
claudecodexPersistent memory system for Claude Code - context compression across sessions
22 skills · 1 MCP server · hooks
Every plugin for OpenAI Codex on GitHub — repos with a .codex-plugin/plugin.json or a Codex marketplace. Install with codex plugin marketplace add owner/repo then codex plugin add name@marketplace. Many work in both tools; see Claude Code plugins.
Persistent memory system for Claude Code - context compression across sessions
22 skills · 1 MCP server · hooks
Claude Code plugin - structured agents with Codex and Claude CLI bridges
12 skills · 10 agents
Local-first, git-native visible-WIP work tracking: plans become tickets, roadmaps stay generated, unplanned work gets recorded. Skills, /worklog:* commands, and an ExitPlanMode capture hook.
15 skills · 5 commands · hooks
Build, migrate, debug, test, secure, and deploy Meteor 3 applications.
20 skills
Preview: one orchestrator session runs work across several repositories or parallel streams of one repository - plan, work packages, dispatch with worktrees and locks, PR review with a read-only reviewer agent, owner queue, closing a program when its goal is reached, state in files. Start with /pepper-orchestrator:init <program>, then /pepper-orchestrator:plan <task>.
1 skill · 13 commands · 3 agents
12 personal skills: motion-design (videos people are proud to share), image-to-frontend, insistir (agents checking agents), goal-loop, second-opinion (GPT-6 Astra as advisor, adversarial reviewer or judge, read-only), remote-agents (SSH), agent-sdk-wizard, managed-agents-wizard, metaprompt, askcodex, compound-knowledge, file-todos
12 skills · 4 agents · hooks
LLPM's product-management and project-management skills: requirements, specs, user stories, stakeholders, meetings, triage, risk, timelines, and dependency/critical-path planning
16 skills
Portable Codex plugin packaging for the public run-project skill bundle.
16 skills
A curated bundle for taking a project from concept to merged work: project setup and analysis, issue decomposition, story-point and change-footprint estimates, user stories, verification plans, architecture/dependency/Mermaid diagrams, and agent-driven work on issues and epics
15 skills
Build and integrate with Esri developer technologies using agent skills for ArcGIS SDKs, APIs, and supported open-source libraries and an MCP server for official documentation.
15 skills
Plugin voor de Nederlandse Richtlijn Digitale Systemen (NeRDS). Bevat skills voor 13 richtlijnen die richting geven aan het ontwerpen, ontwikkelen en inkopen van digitale systemen binnen de Nederlandse overheid: gebruikersbehoeften, toegankelijkheid, open source, open standaarden, cloud, veiligheid, privacy, samenwerking, integratie, data, algoritmen, inkoop en duurzaamheid.
14 skills
Bouw applicaties met het NLDD Designsysteem (@nldd/design-system). Zes skills: nldd-design om componenten, attributen, iconen en releases op te zoeken, nldd-design-build om een applicatie te bouwen, nldd-design-validate om het resultaat te controleren met nldd lint, nldd-design-migrate om een bestaande frontend om te zetten, nldd-design-upgrade om een versie te verhogen, en nldd-design-contribute om een wijziging voor te stellen. Voor consumenten van het pakket, niet voor wie het systeem zelf onderhoudt.
14 skills
Skills to define, apply, and maintain developer design selections.
14 skills
Salesforce compound engineering: ten skills (plan, work, generate, review, debug, deploy, ship, compound, lfg, know) with on-demand modes, stack review lenses, and institutional knowledge compounding for Apex, LWC, Flow, integrations, and Agentforce.
11 skills · 2 MCP servers · hooks
Jam recordings (video, console, network, logs) as context for your coding agent.
13 skills · 1 MCP server
No description.
11 skills · 2 agents
New branch, create/review/merge PR, tag release, morning briefing, end-of-day recap — synced with Jira, Slack, Confluence, GitHub, Gmail and Obsidian
13 skills
Crustdata for Claude: MCP access to 1B+ profiles and 60M+ companies, plus bundled research skills.
9 skills · 2 commands · 1 MCP server · hooks
Skills for omics pipelines, literature discovery, and data visualization.
12 skills · 1 agent
Read a posting, score how well your CV answers it, tailor the CV and write the letter out of facts your profile already holds, render the PDFs, and work out who to message. Applies to nothing. No API key.
12 skills
Teaches agents how to use Dispatch: shared memory, task boards, subagents, repo tools, reaching the user, sharing artifacts, review workflow, UI validation, personas, jobs, templates, and personalities.
12 skills
0.45.8: the recall hint after a failed test run names the exact recall tool and says to call it now, the wording that was followed in every measured Claude Code run (the advisory text was followed in none). 0.45.7: the hint hooks' git-options group is disjoint (a -C or -c value cannot start with a dash), the rule the Hermes plugin now uses; bash's regex was already fast, so this is for parity. 0.45.6: the hint hooks match a git commit after variable assignments with disjoint value pieces, the rule the Hermes plugin now uses (its old rule backtracked exponentially; grep -E never did). 0.45.5: the hint and recall-nudge hooks claim their state with an exclusive file create (bash noclobber, O_EXCL) instead of mkdir, which uutils coreutils on Ubuntu 26.04 does not make exclusive, so a hint fires once, a session's outcomes upload once, and each read takes its own nudge slot (GitHub #687). 0.45.4: the User-Agent hints.sh sends names 0.45.4; the shared hint script is unchanged otherwise, and hooks/hints_cases.json in the Claude Code bundle now holds the hint rules to the same cases as the Hermes plugin's Python port (GitHub #686). 0.45.3: the design-edit hint matches a design document whatever the case of its name, so https://docs.vornik.io counts (GitHub #682). 0.45.2: the recall nudge keeps its state only in the plugin's data directory, or in the per-user runtime directory when the user owns it, never in a shared temp path, and refuses a state directory that is a symlink or that another user owns (GitHub #651). 0.45.1: the design-edit hint also sees a design document written through a shell command (a heredoc, a redirection, tee, sed -i, cp or mv), read from the command text and never run (GitHub #653). 0.45.0: just-in-time hints (GitHub #634). hooks/hints.sh, shared with the Claude Code bundle, prints a short hint from the project's catalogue when it applies: after a commit (send the diff for review), after a design document is patched (ingest it), after a test run whose
10 skills · 1 MCP server · hooks
Skills covering MVUX, navigation, toolkit, themes (Material, Simple, shared semantic), authentication, Kiota API clients, and testing workflows for Uno Platform. The Uno documentation MCP is provided by the Uno tooling, not bundled in the plugin.
9 skills · 1 agent · hooks
Stop prompting. Start pointing. Visual Stack adds a Figma-like feedback layer to Claude Code. Create a new screen from a prompt, screenshot, reference site, or design system, or open an app you already have running. Click any element and leave feedback exactly where the problem is hiding, and Claude publishes the next revision into the same workspace. Compare revisions on a timeline, preview desktop, tablet, and mobile layouts, and keep every comment attached to the element, route, and version it refers to. Wireframes are self-contained HTML and review state stays in your project.
10 skills · hooks
Comprehensive GreyCat development skill and LSP for graph-based language with built-in persistence. Covers data modeling, API development, parallel processing, frontend integration, and all standard libraries.
1 skill · 8 commands · LSP
CONCEPT — Skills voor Nederlandse overheidsstandaarden (beheerd door Logius): API Design Rules, Digikoppeling, OAuth NL, FSC, Logboek Dataverwerkingen, CloudEvents, BOMOS, E-Government en Publicatie-tooling. Bevat 10 skills voor 77 GitHub repositories verdeeld over 9 domeinen.
10 skills
Opinionated skills for planning, auditing, shipping, and improving software.
10 skills
Connect your AI agent to Pachca team chat: the MCP server with a tool per API method, plus nine skills, from summarizing a thread to watching one and replying in it.
9 skills · 1 MCP server
Read-only operations diagnosis, deployment gates, security review, and recovery planning workflows.
9 skills
Bootstraps or standardizes the complete AI-agent harness for a repo - agents, path-scoped rules, commands, hooks, settings.json, docs tree, and AGENTS.md/CLAUDE.md - with a read-only audit mode as well.
1 skill · 8 commands