Struct LearningConfig 

pub struct LearningConfig {
    pub enabled: bool,
    pub min_tool_sequence: usize,
    pub min_success_ratio: f64,
    pub priority_boost_on_success: i32,
    pub priority_decay_on_failure: i32,
    pub max_learned_skills: usize,
    pub stale_procedural_days: u32,
    pub dead_skill_priority_threshold: i64,
}

Configuration for the procedural-memory learning loop.

When sessions close, the governor analyses tool-call sequences and synthesises reusable skill documents for patterns that meet the min_tool_sequence and min_success_ratio thresholds.

Priority asymmetry (intentional)

priority_decay_on_failure (default 10) is 2× larger than priority_boost_on_success (default 5). This deliberate asymmetry ensures that a single failure erases roughly two prior successes, which means unreliable procedures decay to zero and are eventually superseded by better alternatives. Skills start at priority 50, giving a new procedure ten consecutive failures of runway before it drops off the active set (with no successes to counteract).

Fields

enabled: bool

min_tool_sequence: usize

Minimum number of tools in a successful sequence to consider it a procedure.

min_success_ratio: f64

Minimum tool success ratio in a session (0.0–1.0) to trigger learning.

priority_boost_on_success: i32

Priority points added when a learned skill is observed succeeding again.

Default: 5. See struct-level docs for the intentional 2:1 asymmetry with priority_decay_on_failure.

priority_decay_on_failure: i32

Priority points subtracted when a learned skill’s procedure fails.

Default: 10 (2× priority_boost_on_success). The asymmetry causes unreliable skills to decay faster than reliable skills accumulate trust, preventing low-quality procedures from persisting in the active set.

max_learned_skills: usize

Cap on total learned skills to prevent unbounded growth.

stale_procedural_days: u32

Days of zero-activity before a procedural-memory entry is pruned.

Entries in procedural_memory that have never been observed succeeding or failing (success_count = 0 AND failure_count = 0) and are older than this many days are removed during the retrieval-hygiene sweep. Default: 30.

dead_skill_priority_threshold: i64

Learned-skill priority at or below which a skill is considered dead and removed during retrieval-hygiene. Default: 0.

The corresponding .md file on disk is also deleted. Raise this threshold to be more aggressive about culling low-value skills.

Trait Implementations

impl Clone for LearningConfig

fn clone(&self) -> LearningConfig

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for LearningConfig

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

Formats the value using the given formatter.

impl Default for LearningConfig

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for LearningConfig

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for LearningConfig

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

where __S: Serializer,

Serialize this value into the given Serde serializer.