agent-startup-and-context-ingestion

1# Agent Startup and Context Ingestion
2 
3How an agent becomes useful after launch: **AGENTS.md overlays, role
4files, skills, rig specs, workflow specs, startup checklists, refocus
5messages**, and the current `rig context` retrieval and profile surface.
6 
7It is the current concrete startup path inside the broader
8context-engineering-and-retrieval primitive. Startup gets a seat into
9the right initial shape; context engineering is the larger question of
10how a seat gets the right context for the work it is doing right now.
11 
12**Most coordination failures are not tool failures; they are context
13failures.** Agents need to know their role, operating mode, coordination
14convention, boundaries, and current product intent. If startup context
15is scattered or stale, agents execute the wrong thing very efficiently.
16 
17## Use this when
18 
19- Authoring a new agent's startup files (role / culture / startup-context)
20- Refreshing a seat that's been running on stale guidance
21- Auditing whether agents got the current operating mode (not just what's in old files)
22- Designing the orchestrator → next-agent context-transmission shape
23- Building a startup map for OpenRig-building rigs
24 
25## Don't use this when
26 
27- The agent is being created via Agent Starter — the starter's manifest carries startup context
28- The work is artifact-backed mental-model rebuild from a packet — that's `session-compaction-and-restore`
29- The intent is to ship reusable startup content as a skill
30 
31## Failure modes (4)
32…