You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm planning to use Agent OS for a new project where I've already created a comprehensive PRD with:
Detailed interfaces
Specific library choices
Implementation algorithms
Concrete API usage patterns
The Challenge
When I run /create-spec with my detailed PRD as input, Agent OS correctly abstracts this into high-level requirements in technical-spec.md following its templates:
spec.md: 1-2 sentence overview, numbered feature list
technical-spec.md: High-level requirements (e.g., "data model with properties")
This abstraction makes sense for token efficiency and clean planning, but I want the implementation phase to have access to all my PRD details.
My Proposed Workflow
After analyzing the Agent OS instruction files, I noticed that:
/create-tasks generates abstract TDD-structured tasks (1-5 major tasks with test/implementation/verify pattern)
/execute-tasks → execute-task.md Step 2 reads technical-spec.md during implementation
So my proposed workflow would be:
# 1. Create abstract specs
/create-spec (with detailed PRD.md as input)
# 2. Generate abstract tasks
/create-tasks
# 3. Inject detailed PRD before execution
cp my-detailed-PRD.md .agent-os/specs/.../sub-specs/technical-spec.md
# 4. Execute with full details
/execute-tasks
Rationale:
Tasks stay abstract and flexible (TDD pattern intact)
Implementation phase gets all concrete details from technical-spec.md
Best of both worlds: clean planning + detailed execution
Questions
Is this the recommended approach when working with pre-existing detailed PRDs?
Are there any downsides to replacing the abstracted technical-spec.md before /execute-tasks?
Is there a better way to inject detailed specifications into the Agent OS workflow?
I suspect this is a common scenario: sometimes you start with rough ideas (where Agent OS abstraction is perfect), but sometimes you already have a detailed PRD and want that precision during
implementation.
Any guidance from the community or maintainers would be greatly appreciated!
Additional Context:
I've used Agent OS several times before successfully
This is the first time I'm trying to integrate an existing detailed PRD
I analyzed .agent-os/instructions/core/execute-task.md Step 2 which confirms it reads technical-spec.md during execution
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
Context
I'm planning to use Agent OS for a new project where I've already created a comprehensive PRD with:
The Challenge
When I run
/create-spec
with my detailed PRD as input, Agent OS correctly abstracts this into high-level requirements intechnical-spec.md
following its templates:spec.md
: 1-2 sentence overview, numbered feature listtechnical-spec.md
: High-level requirements (e.g., "data model with properties")This abstraction makes sense for token efficiency and clean planning, but I want the implementation phase to have access to all my PRD details.
My Proposed Workflow
After analyzing the Agent OS instruction files, I noticed that:
/create-tasks
generates abstract TDD-structured tasks (1-5 major tasks with test/implementation/verify pattern)/execute-tasks
→execute-task.md
Step 2 reads technical-spec.md during implementationSo my proposed workflow would be:
Rationale:
Questions
I suspect this is a common scenario: sometimes you start with rough ideas (where Agent OS abstraction is perfect), but sometimes you already have a detailed PRD and want that precision during
implementation.
Any guidance from the community or maintainers would be greatly appreciated!
Additional Context:
Beta Was this translation helpful? Give feedback.
All reactions