Skip to content

instruction

Instruction components: rules, memory and textual constraints for a harness.

instruction

instruction is the textual part of a setup: rules for how an agent behaves, project memory, working style, the limits of its authority, and hints for one particular harness.

An instruction runs no code and connects no external tool. Its power is that it changes the context an agent decides in.

When to use it

SituationIs instruction rightWhy
"Always write commit messages in English"yesit is a rule of behaviour
"Read the testing guide before backend tests"yesit routes attention
"Run the scanner before push"partlythe rule can be written down, but the running belongs in a command or a hook
"Connect the GitHub MCP"nothat is a separate mcp
"Store the API token"nosecrets do not go into a component

How it works in ai_stp

  1. An author publishes or imports an instruction with a passport.
  2. The passport records the source, the version, the supported harnesses and the scope.
  3. The setup compiler checks the instruction against the chosen harness.
  4. The provider projects the text onto the native surface: an instruction file, a memory surface, or another supported format.

!!! note "Memory is not a component kind" Memory, rules, preferences and project conventions are the content of an instruction or a setting. There is no memory kind in ai_stp.

What gets recorded

What is recordedWhy
Provenanceto know who supplied the text, from which commit and path
Versionto tell an immutable X.Y release from the next one
Harnessso Claude-specific text is not applied to an incompatible target
Scopeto limit a rule to a project, a user or one setup
Trust lineto decide whether it may be shown or installed without manual consent

??? warning "The main risk" An instruction can be harmless text, and it can quietly widen an agent's authority. Before installing, read the diff of the content, not only the title.

ai_stp