Framework Differentiation¶
ORCA should be differentiated by the work it is built to support and the evidence it chooses to trust.
What ORCA Is Not¶
ORCA is not mainly:
- a prompt library
- a generic AI-agent wrapper
- a note-taking layer with no delivery path
- a collection of unrelated commands with no default workflow
What Actually Makes ORCA Different¶
It Treats Workflow As The Product¶
ORCA is built around the places real work breaks:
- vague starts
- missing specs
- lost context between stages
- skipped review and QA
- no receipt for what happened
- weak follow-up on blockers or next actions
It Stays Close To Real Operating Context¶
ORCA is designed for work that spans:
- app shipping and release operations
- research capture and staged promotion
- blocker tracking and stale-work cleanup
- mixed harness, repo, and tool environments
It Uses Stronger Sources Of Truth¶
ORCA is supposed to trust:
- the vault when available
- the repo and its artifacts
- the declared system of record
before it falls back to generic outside advice.
It Teaches Without Becoming A Tutorial-Only Layer¶
ORCA is meant to onboard new users into orchestration while remaining a durable long-term operating surface.
That is different from systems that feel like a beginner wrapper before users graduate to something else.
Good Differentiation Language¶
Prefer language like:
- durable workflow
- inspectable execution
- blocker-aware delivery
- staged research to action flow
- receipts, lineage, and review
- one clear default path with deeper layers behind it
Bad Differentiation Language¶
Avoid language like:
- revolutionary AI operating system
- universal agent platform for every workflow
- seamless autonomous intelligence layer
- anything that sounds detached from the actual repo and vault evidence