Struct SkillsConfig 

pub struct SkillsConfig {

    pub paths: Vec<String>,
    pub max_active_skills: usize,
    pub disambiguation_threshold: f32,
    pub min_injection_score: f32,
    pub cosine_weight: f32,
    pub hybrid_search: bool,
    pub learning: LearningConfig,
    pub trust: TrustConfig,
    pub prompt_mode: SkillPromptMode,
    pub two_stage_matching: bool,
    pub confusability_threshold: f32,
    pub rl_routing_enabled: bool,
    pub rl_learning_rate: f32,
    pub rl_weight: f32,
    pub rl_persist_interval: u32,
    pub rl_warmup_updates: u32,
    pub rl_embed_dim: Option<usize>,
    pub generation_provider: ProviderName,
    pub generation_output_dir: Option<String>,
    pub mining: SkillMiningConfig,
}

Skill discovery and matching configuration, nested under [skills] in TOML.

Controls where skills are loaded from, how they are ranked during retrieval, the RL re-ranking head, NL skill generation, and automated skill mining.

Example (TOML)

[skills]
paths = ["~/.config/zeph/skills"]
max_active_skills = 5
disambiguation_threshold = 0.20
hybrid_search = true

Fields

paths: Vec<String>

Directories to scan for *.skill.md / SKILL.md files.

max_active_skills: usize

disambiguation_threshold: f32

min_injection_score: f32

cosine_weight: f32

hybrid_search: bool

learning: LearningConfig

trust: TrustConfig

prompt_mode: SkillPromptMode

two_stage_matching: bool

Enable two-stage category-first skill matching (requires category set in SKILL.md). Falls back to flat matching when no multi-skill categories are available.

confusability_threshold: f32

Warn when any two skills have cosine similarity ≥ this threshold. Set to 0.0 (default) to disable the confusability check entirely.

rl_routing_enabled: bool

Enable RL routing head for skill re-ranking (disabled by default).

rl_learning_rate: f32

Learning rate for REINFORCE weight updates.

rl_weight: f32

Blend weight: final_score = (1-rl_weight)*cosine + rl_weight*rl_score.

rl_persist_interval: u32

Persist weights every N updates (0 = persist every update).

rl_warmup_updates: u32

Skip RL blending for the first N updates (cold-start warmup).

rl_embed_dim: Option<usize>

Embedding dimension for the RL routing head. Must match the output dimension of the configured embedding provider. Defaults to None → 1536 (text-embedding-3-small output dimension).

generation_provider: ProviderName

Provider name for /skill create NL generation. Empty = primary provider.

generation_output_dir: Option<String>

Directory where generated skills are written. Defaults to first entry in paths.

mining: SkillMiningConfig

Skill mining configuration.

Trait Implementations

impl Debug for SkillsConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for SkillsConfig

fn deserialize<__D>( __deserializer: __D, ) -> Result<SkillsConfig, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for SkillsConfig

fn serialize<__S>( &self, __serializer: __S, ) -> Result<<__S as Serializer>::Ok, <__S as Serializer>::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.