Agently Skills
Nine skills from the Agently framework authors: task routing, model requests, TriggerFlow workflows and migration from LangChain
- 35 ★ on the repository
- Last commit
Minimal access
Why this level
- The skills only instruct the agent and run nothing themselves
- The Codex installer writes files into the skills folder and keeps backups
Install
- Works in
Manual install
npx skills add AgentEra/Agently-Skills --agent "$AGENT" --skill agently -yRun `export AGENT=codex` first. This command installs the single router skill.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills only instruct the agent and run nothing themselves
- The Codex installer writes files into the skills folder and keeps backups
This is third-party code. Review the repository files before installing.
What it does
The AgentEra repository holds nine skills for agents that write applications on the Agently framework, plus one deprecated entry. The entry point `agently` analyzes a task and routes it to the right specialist. `agently-request` handles model requests, prompts, structured output, session memory, embeddings and retrieval. `agently-runtime` covers Action Runtime, MCP, TaskWorkspace and RecordStore. `agently-triggerflow` builds branching, concurrent and resumable workflows with inspectable state. `agently-stage` deals with task lifetime, streams, events and backpressure. `agently-agentic-design` designs and audits non-trivial systems across owner layers. `agently-migration` maps LangChain, LangGraph, LlamaIndex and CrewAI onto Agently's native layers. `agently-product-design` and `agently-collaboration` have no framework dependency: the first helps with product and document design, the second with progress reports and handoffs. The v3 catalog matches Agently 4.1.4.10.
Who it is for. Developers building on Agently and teams migrating projects from LangChain or CrewAI.
Good fit when
- A new Agently application starts with the question of who owns what: model, host or workflow
- You need structured output and session memory for model requests
- A process has to branch, run in parallel and resume after a restart
- A LangGraph or CrewAI project is being moved to Agently layer by layer
- An agent works through a long task and you want a progress board and recorded decisions
Not a fit when
- The project does not use Agently: the layer-specific skills will not help, only product-design and collaboration apply
- You stay on another stack: the `agently` skill respects the chosen framework by default and will not port code to Agently
- Your Agently is older than 4.1.4.10: the v1 and v2 archive branches exist for rollback, not for new installs
Example request
Design an Agently assistant that triages requests, calls tools and resumes after a failure. First say which layer owns what.Limitations
The v3 catalog is aligned with Agently 4.1.4.10. Earlier generations are frozen on the `update/archive-v2-catalog` and `update/archive-legacy-v1-catalog` branches for rollback.
The `agently-design` entry is deprecated and forwards to `agently-agentic-design`.
The `npx skills add` commands are generic for any repository with a SKILL.md, and for Codex the project has its own installer, `scripts/install.py`, on Python 3.10 with PyYAML.
The installer does not edit the agent's global instructions: the guidance block for `agently-collaboration` is added by hand.
How to disable. Delete the `agently*` skill directories you installed from the agent's skills folder. For Codex that is `$CODEX_HOME/skills` or `~/.codex/skills`, with backups beside it in `skill-update-backups/`.
FAQ
Which skill do I start with?
With `agently`: it picks the specialist itself. Installing only this skill gives the smallest discovery surface.
How do I install all the main skills?
Pick an agent with `export AGENT=codex`, then loop `npx skills add AgentEra/Agently-Skills --agent "$AGENT" --skill <name> -y` over the seven app skills: `agently`, `agently-collaboration`, `agently-agentic-design`, `agently-request`, `agently-runtime`, `agently-stage`, `agently-triggerflow`.
Do I need Agently for product design?
No, `agently-product-design` and `agently-collaboration` work without the framework. They can be installed alone, for example with `python scripts/install.py --skills agently-collaboration`.
How do I change the skills' interface language in Codex?
Run `python scripts/install.py --locale en --ui-only` for English or `--locale zh-CN` for Chinese. The language is picked at install time, not by the conversation language.
What should I do with the agently-design skill?
Replace it with `agently-agentic-design`. The old entry only forwards to it; on the next update check the callers and remove it.
Discussion
No questions or reviews yet.
Related
SkillsA catalog of nearly two hundred skills: code delivery, review, debugging, AI governance, product work and agent coordination.
FastAgent turns an agent directory with skills and TypeScript tools into a live service with bots and schedules
A tool that builds tested skills from a real workflow and governs their lifecycle through your own marketplace
A skill with 25 commands to audit and improve the prompts, tools and agent designs of a project in Cursor, Claude Code and Codex
Official LangChain skills: choose between LangChain, LangGraph and Deep Agents, quickstarts, RAG, memory, human approval and eval engineering
A scaffold of personas, rules and skills in plain Markdown: Maestro hands work to different models and spares your main plan
README in short
The README describes the current nine-skill catalog, compatibility with Agently versions and archive branches. It then covers installation: a Codex installer with an interface language choice (English and Chinese), generic `npx skills add` commands and bundles for apps and migration.
README badge
Original SKILL.md
--- name: agently description: Route ambiguous product, document, or Agentic work to the right design or implementation skill. Use for unresolved owner/project shape, mixed specialties or explicit low-frequency TaskDAG or DynamicTask requests. Clear tasks use their specialist directly; conversation expression, progress and handoffs use agently-collaboration alongside any task. --- # Agently Start here when the owner or working method is unclear, or several specialties share one work item. First distinguish human product/document outcomes from model-service behavior. Route clear product work directly to `agently-product-design`; no model, Agently runtime or framework migration is required. For model systems, choose the smallest suitable owner before files or mechanisms; use Agently-native mappings only when Agently is in scope. Respect another selected stack. When an Agently developer needs to understand model-node duties, tune execution effects, check data flow or redundant context, or judge model/Host ownership, use collaborative review without waiting for the phrase "Prompt review". Route through `agently-agentic-design` for the highlighted flow and handoffs, and `agently-request` for related Prompt tables grouped by readability and dependency. Keep unrelated mechanical work outside this review method. ## Shared coordination Use [Conversation and Collaboration](../agently-collaboration/SKILL.md) alongside any task for reader-oriented communication, progress, decisions and handoffs. It can be selected directly without this router. Specialist Skills retain their execution and acceptance criteria; reuse one work record across them. If the companion is unavailable, follow the host's base guidance and existing task record; routing does not require installation or a new tracking system. ## Workflow The following owner workflow applies to model-powered systems. For non-model product or writing tasks, route to the product Skill and continue its method. 1. Reduce the request to goals, inputs, outputs, side effects, evidence, and lifecycle needs. 2. Separate model-owned semantic decisions from host-owned validation, authorization, identity joins, arithmetic, and effects. 3. Assign each decision, state, and effect to one existing owner. 4. For non-trivial flows, map nodes, value/event edges, invariants, and the production need for every requested field and stage. 5. Implement async-first where the caller owns an async boundary; validate deterministic contracts separately from model-owned quality. 6. Report a framework gap when no native owner can carry a required invariant. Do not hide the gap behind business-specific glue or a pass-through facade. ## Route by Owner - Product goals, role/behavior coverage, interface/component/layout design,