Skip to content

Commit ddc4c25

Browse files
authored
Merge pull request #232 from buildermethods/v2.1
Version 2.1.0
2 parents f6d1477 + a28c9af commit ddc4c25

File tree

62 files changed

+1998
-2867
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

62 files changed

+1998
-2867
lines changed

CHANGELOG.md

Lines changed: 106 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -5,6 +5,112 @@ All notable changes to Agent OS will be documented in this file.
55
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
66
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
77

8+
## [2.1.0] - 2025-10-21
9+
10+
Version 2.1 implemented a round of significant changes to how things work in Agent OS. Here is a summary of what's new in version 2.1.0:
11+
12+
### TL;DR
13+
14+
Here's the brief overview. It's all detailed below and the [docs](https://buildermethods.com/agent-os) have been updated to reflect all of this.
15+
16+
- Option to leverage Claude Code's new "Skills" feature for reading standards
17+
- Option to enable or disable delegating to Claude Code subagents
18+
- Replaced "single/multi-agent modes" with more flexible configuration options
19+
- Retired the short-lived "roles" system. Too complex, and better handled with standard tooling (more below).
20+
- Removed documentation & verification bloat
21+
- Went from 4 to 6 more specific development phases (use 'em all or pick and choose!):
22+
1. plan-product -- (no change) Plan your product's mission & roadmap
23+
2. shape-spec -- For shaping and planning a feature before writing it up
24+
3. write-spec -- For writing your spec.md
25+
4. create-tasks -- For creating your tasks.md
26+
5. implement-tasks -- Simple single-agent implementation of tasks.md
27+
6. orchestrate-tasks -- For more advanced, fine-grain control and multi-agent orchestration of tasks.md.
28+
- Simplified & improved project upgrade script
29+
30+
Let's unpack these updates in detail:
31+
32+
### Claude Code Skills support
33+
34+
2.1 adds official support for [Claude Code Skills](https://docs.claude.com/en/docs/claude-code/skills).
35+
36+
When the config option standards_as_claude_code_skills is true, this will convert all of your standards into Claude Code Skills and _not_ inject references to those Standards like Agent OS normally would.
37+
38+
2.1 also provides a Claude Code command, `improve-skills` which you **definitely should** run after installing Agent OS in your project with the skills option turned on. This command is designed to improve and rewrite each of your Claude Code Skills descriptions to make them more useable and discoverable by Claude Code.
39+
40+
### Enable/Disable delegation to Claude Code subagents
41+
42+
2.1 introduces an config option to enable or disable delegating tasks to Claude Code subagents. You can disable subagents by setting use_claude_code_subagents to false.
43+
44+
When set to false, and when using Claude Code, you can still run Agent OS commands in Claude Code, and instead of delegating most tasks to subagents, Claude Code's main agent will execute everything.
45+
46+
While you lose some context efficiency of using subagents, you can token efficiency and some speed gains without the use of subagents.
47+
48+
### Replaced "single-agent & multi-agent modes" with new config options
49+
50+
2.0.x had introduced the concepts of multi-agent and single-agent modes, where multi-agent mode was designed for using Claude Code with subagents. This naming and configuration design proved suboptimal and inflexible, so 2.1.0 does away with the terms "single-agent mode" and "multi-agent mode".
51+
52+
Now we configure Agent OS using these boolean options in your base ~/agent-os/config.yml:
53+
54+
claude_code_commands: true/false
55+
use_claude_code_subagents: true/false
56+
agent_os_commands: true/false
57+
58+
The benefits of this new configuration approach are:
59+
60+
- Now you can use Agent OS with Claude Code *with* or *without* delegating to subagents. (subagents bring many benefits like context efficiency, but also come with some tradeoffs—higher token usage, less transparency, slower to finish tasks).
61+
62+
- Before, when you had *both* single-agent and multi-agent modes enabled, your project's agent-os/commands/ folder ended up with "multi-agent/" and "single-agent/" subfolders for each command, which is confusing and clumsy to use. Now in 2.1.0, your project's agent-os/commands/ folder will not have these additional "modes" subfolders.
63+
64+
- Easier to integrate additional feature configurations as they become available, so that you can mix and match the exact set of features that fit your preferred coding tools and workflow style. For example, we're also introducing an option to make use of the new Claude Code Skills feature (or you can opt out). More on this below.
65+
66+
### Retired (short-lived) "Roles" system
67+
68+
2.0.x had introduced a concept of "Roles", where your roles/implementers.yml and roles/verifiers.yml contained convoluted lists of agents that could be assigned to implement tasks. It also had a script for adding additional "roles".
69+
70+
All of that is removed in 2.1.0. That system added no real benefit over simply using available tooling (like Claude Code's own subagent generator) for spinning up your subagents.
71+
72+
2.1.0 introduces an 'orchestrate-tasks' phase, which achieves the same thing that the old "Roles" system intended: Advanced orchestration of multiple specialized subagents to carry out a complex implementation. More on this below.
73+
74+
### Removed documentation & verification bloat
75+
76+
2.0.x had introduced a bunch of "bloat" that quickly proved unnecessary and inefficient. These bits have been removed in 2.1.0:
77+
78+
- Verification of your spec (although the spec-verifier Claude Code subagent is still available for you to call on, if/when you want)
79+
- Documentation of every task's implementation
80+
- Specialized verifiers (backend-verifier, frontend-verifier)
81+
82+
The final overall verification step for a spec's implementation remains intact.
83+
84+
### From 4 to 6 more specific development phases
85+
86+
While some users use all of Agent OS' workflow for everything, many have been picking the parts they find useful and discarding those that don't fit their workflow—AS THEY SHOULD!
87+
88+
2.1.0 establishes this as a core principle of Agent OS: You can use as much or as little of it as you want!
89+
90+
With that in mind, we've moved from 4 to 6 different phases of development that can _potentially_ be powered by Agent OS:
91+
92+
1. `plan-product` -- No changes here. This is for establishing your product's mission, roadmap and tech-stack.
93+
94+
2. `shape-spec` -- Use this when you need to take your rough idea for a feature and shape it into a well-scoped and strategized plan, before officially writing it up. This is where the agent asks you clarifying questions and ends up producing your requirements.md.
95+
- Already got your requirements shaped? Skip this and drop those right into your spec's requirements.md 👍
96+
97+
3. `write-spec` -- Takes your requirements.md and formalizes it into a clear and concise spec.md.
98+
99+
4. `create-tasks` -- Takes your spec.md and breaks it down into a tasks list, grouped, prioritized and ready for implementation.
100+
101+
5. `implement-tasks` -- Just want to build right now(!), then use this to implement your tasks.md with your main agent.
102+
103+
6. `orchestrate-tasks` -- Got a big complex feature and want to orchestrate multiple agents, with more fine-grain control over their contexts? Use this. It provides a structure to delegate your task groups to any Claude Code subagents you've created. Or if you're not using Claude Code, it generates targeted prompt files (as was established in 2.0.x).
104+
105+
### Simplified & improved project upgrade script
106+
107+
Now whenever you need to upgrade your Agent OS project installation (to a new version or to push configuration changes or standards changes to a project), now when you run project-install.sh or project-update.sh, the system will:
108+
109+
- Check and compare your incoming version & configs to your current project's
110+
- Show you what will stay intact or be removed & re-installed
111+
- Ask you to confirm to proceed.
112+
113+
8114
## [2.0.5] - 2025-10-16
9115

10116
- Updated base installation update options to include a "Full update" option, which is the easiest way to pull and update the latest Agent OS stuff (default profile, scripts) without losing your base installation's custom profiles.

config.yml

Lines changed: 33 additions & 21 deletions
Original file line numberDiff line numberDiff line change
@@ -1,44 +1,56 @@
1-
version: 2.0.5
1+
version: 2.1.0
22
base_install: true
33

4+
5+
# CONFIGURATION
6+
# Configure defaults for the tools you use and how Agent OS should compile commands for your projects.
7+
8+
49
# ================================================
5-
# MULTI-AGENT MODE
10+
# Do you use Claude Code?
11+
# Set to true to install commands in your project's .claude/commands/agent-os/ folder
612
#
7-
# Use this mode if you're using a tool that supports multiple agents (a.k.a. subagents).
8-
# In multi-agent mode, you initiate commands (new-spec, create-spec, implement-spec, etc.)
9-
# your main agent orchestrates multiple subagents to perform long-running tasks autonomously.
10-
#
11-
# You can override this by running:
12-
# ~/.agent-os/scripts/project-install.sh --multi-agent-mode [true/false] --multi-agent-tool [tool-name]
13+
# Override this default when running project-install.sh by using the flag --claude-code-commands=true/false
1314
# ================================================
15+
claude_code_commands: true
1416

15-
multi_agent_mode: true
1617

17-
# Currently supported tools: claude-code
18-
multi_agent_tool: claude-code
18+
# ================================================
19+
# Do you use other coding tools (Cursor, Windsurf, etc.)?
20+
# Set to true to install commands in your project's agent-os/commands/ folder
21+
#
22+
# Override this default when running project-install.sh by using the flag --agent-os-commands true/false
23+
# ================================================
24+
agent_os_commands: false
1925

2026

2127
# ================================================
22-
# SINGLE-AGENT MODE
23-
#
24-
# Use this mode if you're using a tool that doesn't support multiple agents (a.k.a. subagents).
25-
# In single-agent mode, you deliver prompts to your main agent to perform tasks step-by-step.
28+
# Do you want Claude Code to use subagents?
29+
# Set to true to install agents in .claude/agents/agent-os/ and have commands delegate to them
30+
# Requires claude_code_commands: true
2631
#
27-
# You can override this by running:
28-
# ~/.agent-os/scripts/project-install.sh --single-agent-mode [true/false] --single-agent-tool [tool-name]
32+
# Override this default when running project-install.sh by using the flag --use-claude-code-subagents true/false
2933
# ================================================
34+
use_claude_code_subagents: true
35+
3036

31-
single_agent_mode: false
32-
single_agent_tool: generic
37+
# ================================================
38+
# Should standards be provided to Claude Code as Skills?
39+
# Set to true to use Claude Code's Skills feature for reading standards
40+
# Set to false to inject standards as file references in command prompts
41+
# Requires claude_code_commands: true (automatically treated as false if claude_code_commands is false)
42+
#
43+
# Override this default when running project-install.sh by using the flag --standards-as-claude-code-skills true/false
44+
# ================================================
45+
standards_as_claude_code_skills: true
3346

3447

3548
# ================================================
3649
# PROFILE
3750
#
3851
# Use this profile by default.
3952
#
40-
# You can override this by running:
41-
# ~/.agent-os/scripts/project-install.sh --profile [profile-name]
53+
# Override this default when running project-install.sh by using the flag --profile [profile-name]
4254
# ================================================
4355

4456
profile: default
Lines changed: 7 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
---
22
name: implementation-verifier
3-
description: Verify the end-to-end implementation of a spec
3+
description: Use proactively to verify the end-to-end implementation of a spec
44
tools: Write, Read, Bash, WebFetch, Playwright
55
color: green
66
model: inherit
@@ -11,29 +11,24 @@ You are a product spec verifier responsible for verifying the end-to-end impleme
1111
## Core Responsibilities
1212

1313
1. **Ensure tasks.md has been updated**: Check this spec's `tasks.md` to ensure all tasks and sub-tasks have been marked complete with `- [x]`
14-
2. **Verify that implementations and verifications have been documented**: Ensure this spec's `implementation/` and `verification` folders contain documentation from each implementer and verifier.
15-
3. **Update roadmap (if applicable)**: Check `agent-os/product/roadmap.md` and check items that have been completed as a result of this spec's implementation by marking their checkbox(s) with `- [x]`.
16-
4. **Run entire tests suite**: Verify that all tests pass and there have been no regressions as a result of this implementation.
17-
5. **Create final verification report**: Write your final verification report for this spec's implementation.
14+
2. **Update roadmap (if applicable)**: Check `agent-os/product/roadmap.md` and check items that have been completed as a result of this spec's implementation by marking their checkbox(s) with `- [x]`.
15+
3. **Run entire tests suite**: Verify that all tests pass and there have been no regressions as a result of this implementation.
16+
4. **Create final verification report**: Write your final verification report for this spec's implementation.
1817

1918
## Workflow
2019

2120
### Step 1: Ensure tasks.md has been updated
2221

2322
{{workflows/implementation/verification/verify-tasks}}
2423

25-
### Step 2: Verify that implementations and verifications have been documented
26-
27-
{{workflows/implementation/verification/verify-documentation}}
28-
29-
### Step 3: Update roadmap (if applicable)
24+
### Step 2: Update roadmap (if applicable)
3025

3126
{{workflows/implementation/verification/update-roadmap}}
3227

33-
### Step 4: Run entire tests suite
28+
### Step 3: Run entire tests suite
3429

3530
{{workflows/implementation/verification/run-all-tests}}
3631

37-
### Step 5: Create final verification report
32+
### Step 4: Create final verification report
3833

3934
{{workflows/implementation/verification/create-verification-report}}
Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
---
2+
name: implementer
3+
description: Use proactively to implement a feature by following a given tasks.md for a spec.
4+
tools: Write, Read, Bash, WebFetch, Playwright
5+
color: red
6+
model: inherit
7+
---
8+
9+
You are a full stack software developer with deep expertise in front-end, back-end, database, API and user interface development. Your role is to implement a given set of tasks for the implementation of a feature, by closely following the specifications documented in a given tasks.md, spec.md, and/or requirements.md.
10+
11+
{{workflows/implementation/implement-tasks}}
12+
13+
{{UNLESS standards_as_claude_code_skills}}
14+
## User Standards & Preferences Compliance
15+
16+
IMPORTANT: Ensure that the tasks list you create IS ALIGNED and DOES NOT CONFLICT with any of user's preferred tech stack, coding conventions, or common patterns as detailed in the following files:
17+
18+
{{standards/*}}
19+
{{ENDUNLESS standards_as_claude_code_skills}}

profiles/default/agents/product-planner.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
---
22
name: product-planner
3-
description: Create product documentation including mission, and roadmap
3+
description: Use proactively to create product documentation including mission, and roadmap
44
tools: Write, Read, Bash, WebFetch
55
color: cyan
66
model: inherit
@@ -53,8 +53,10 @@ done
5353
echo "Product planning complete! Review your product documentation in agent-os/product/"
5454
```
5555

56+
{{UNLESS standards_as_claude_code_skills}}
5657
## User Standards & Preferences Compliance
5758

5859
IMPORTANT: Ensure the product mission and roadmap are ALIGNED and DO NOT CONFLICT with the user's preferences and standards as detailed in the following files:
5960

6061
{{standards/global/*}}
62+
{{ENDUNLESS standards_as_claude_code_skills}}

profiles/default/agents/specification/spec-initializer.md renamed to profiles/default/agents/spec-initializer.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
---
22
name: spec-initializer
3-
description: Initialize spec folder and save raw idea
3+
description: Use proactively to initialize spec folder and save raw idea
44
tools: Write, Bash
55
color: green
66
model: sonnet
Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
---
2-
name: spec-researcher
3-
description: Gather detailed requirements through targeted questions and visual analysis
2+
name: spec-shaper
3+
description: Use proactively to gather detailed requirements through targeted questions and visual analysis
44
tools: Write, Read, Bash, WebFetch
55
color: blue
66
model: inherit
@@ -10,8 +10,10 @@ You are a software product requirements research specialist. Your role is to gat
1010

1111
{{workflows/specification/research-spec}}
1212

13+
{{UNLESS standards_as_claude_code_skills}}
1314
## User Standards & Preferences Compliance
1415

1516
IMPORTANT: Ensure that all of your questions and final documented requirements ARE ALIGNED and DO NOT CONFLICT with any of user's preferred tech-stack, coding conventions, or common patterns as detailed in the following files:
1617

1718
{{standards/*}}
19+
{{ENDUNLESS standards_as_claude_code_skills}}

profiles/default/agents/specification/spec-verifier.md renamed to profiles/default/agents/spec-verifier.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
---
22
name: spec-verifier
3-
description: Verify the spec and tasks list
3+
description: Use proactively to verify the spec and tasks list
44
tools: Write, Read, Bash, WebFetch
55
color: pink
66
model: sonnet
@@ -10,8 +10,10 @@ You are a software product specifications verifier. Your role is to verify the s
1010

1111
{{workflows/specification/verify-spec}}
1212

13+
{{UNLESS standards_as_claude_code_skills}}
1314
## User Standards & Preferences Compliance
1415

1516
IMPORTANT: Ensure that the spec and tasks list are ALIGNED and DO NOT CONFLICT with any of user's preferred tech stack, coding conventions, or common patterns as detailed in the following files:
1617

1718
{{standards/*}}
19+
{{ENDUNLESS standards_as_claude_code_skills}}

profiles/default/agents/specification/spec-writer.md renamed to profiles/default/agents/spec-writer.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
---
22
name: spec-writer
3-
description: Create a detailed specification document for development
3+
description: Use proactively to create a detailed specification document for development
44
tools: Write, Read, Bash, WebFetch
55
color: purple
66
model: inherit
@@ -10,8 +10,10 @@ You are a software product specifications writer. Your role is to create a detai
1010

1111
{{workflows/specification/write-spec}}
1212

13+
{{UNLESS standards_as_claude_code_skills}}
1314
## User Standards & Preferences Compliance
1415

1516
IMPORTANT: Ensure that the spec you create IS ALIGNED and DOES NOT CONFLICT with any of user's preferred tech stack, coding conventions, or common patterns as detailed in the following files:
1617

1718
{{standards/*}}
19+
{{ENDUNLESS standards_as_claude_code_skills}}
Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,17 +1,19 @@
11
---
22
name: task-list-creator
3-
description: Create a detailed and strategic tasks list for development of a spec
3+
description: Use proactively to create a detailed and strategic tasks list for development of a spec
44
tools: Write, Read, Bash, WebFetch
55
color: orange
66
model: inherit
77
---
88

99
You are a software product tasks list writer and planner. Your role is to create a detailed tasks list with strategic groupings and orderings of tasks for the development of a spec.
1010

11-
{{workflows/specification/create-tasks-list}}
11+
{{workflows/implementation/create-tasks-list}}
1212

13+
{{UNLESS standards_as_claude_code_skills}}
1314
## User Standards & Preferences Compliance
1415

1516
IMPORTANT: Ensure that the tasks list you create IS ALIGNED and DOES NOT CONFLICT with any of user's preferred tech stack, coding conventions, or common patterns as detailed in the following files:
1617

1718
{{standards/*}}
19+
{{ENDUNLESS standards_as_claude_code_skills}}

0 commit comments

Comments
 (0)