mirror of
https://github.com/openai/codex.git
synced 2026-11-13 06:45:07 +00:00
jif
c100109280
## What changed - Add an opt-in `skill_search` feature that ranks prompt-visible skills against each turn's user input with a bounded weighted lexical selector. - Keep the ranked selection out of model-visible context and record metrics for selection cost, catalog reduction, and whether later implicit or `skills.read` invocations matched the ranked candidates. - Include host-provided skills in the experiment catalog without changing the rendered skill catalog. ## Testing - Add selector unit tests covering ranking, limits, truncation, stop words, and deterministic tie-breaking. - Add extension tests covering turn-local invocation recording and host-skill shadow selection. GitOrigin-RevId: 4d00a1c805ea8b391d6c6ac6a8450afa88ca3e25
13 lines
590 B
Rust
13 lines
590 B
Rust
/// Host-supplied configuration used by the skills extension.
|
|
#[derive(Clone, Debug, Eq, PartialEq)]
|
|
pub struct SkillsExtensionConfig {
|
|
/// Whether the available-skills catalog is included in model context.
|
|
pub include_instructions: bool,
|
|
/// Whether bundled skills are eligible for discovery.
|
|
pub bundled_skills_enabled: bool,
|
|
/// Whether orchestrator-owned skills are eligible for discovery.
|
|
pub orchestrator_skills_enabled: bool,
|
|
/// Whether cheap skill selectors run in shadow mode without changing prompt contents.
|
|
pub shadow_selection_enabled: bool,
|
|
}
|