Skip to content

Memory and agent instructions

Agent instructions set shared working rules, and SASE memory holds context those instructions and agents can reuse.

The idea in one picture

Shared instructions and memory ──► Agent turn ◄── Task prompt
          project or home                         this task

The diagram separates enduring context from the prompt for the current task.

How it works

An instruction file states rules that should outlast one prompt. Memory stores durable context that agents can share across turns, such as project conventions or reusable reference notes. Some memory is loaded as instruction context; reference material can be consulted when a task needs it.

The task prompt then says what the agent should work on now. Keeping shared expectations in instructions and memory, and the immediate goal in the prompt, avoids rewriting the same context for every task.

When it matters

  • Put stable project expectations in shared instructions so each agent starts with the same boundaries.
  • Keep reusable knowledge in memory when it should remain available beyond one conversation.
  • Use the prompt for the work that is specific to this turn.

Do it

Look it up

  • Memory — memory types and how to inspect them.
  • Instruction bundles — how instructions are assembled and delivered.
  • Glossary — agent instructions, core memory, and reference memory.
  • Design decisions — why provider adapters deliver instructions per invocation.

_[project]: A sase project is a named unit of work registered with SASE. _[workspace]: A sase workspace is a numbered clone of a project's primary repo, managed by the workspace store and tracked in that project's registry.json. _[agent]: A sase agent is an agent session or a single agent that does not belong to a session. _[prompt]: A prompt is the text SASE hands to a provider to start an agent turn. _[macro]: A macro is a reusable prompt triggered with #name in agent prompts. _[Patch]: A Patch is SASE's local unit of change. _[bead]: A bead is one tracked work item in SASE's git-native issue tracker. _[plan]: A plan is a persisted, approved artifact recording the intent behind agent work. _[TUI]: The TUI is SASE's terminal interface for running agents and reviewing results. _[epic]: An epic is an executable multi-phase plan run with sase bead work. _[gate]: A sase gate is a durable, command-backed request for one user decision, created when work must pause for the user. _[monitor]: A sase monitor is a monitor turn that runs one long command under a detached supervisor, so the command outlives the agent that started it. *[stitch]: A stitch is the lightweight ordered change record inside a Patch's STITCHES: section.