Documentation
¶
Overview ¶
Package values models the moral compass and ethical guidelines of an agent.
Values are the rules and principles that govern what is acceptable. Toxicity detection is a values question: "Does this violate our ethical standards?" Brand safety is a values question: "Does this align with our brand principles?" Values evaluate content against principles with configurable thresholds and severity levels.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type ListFilter ¶
ListFilter defines filtering options for listing values.
type Principle ¶
type Principle string
Principle defines a specific ethical or moral rule category.
const ( // PrincipleToxicity evaluates for toxic, harmful, or hateful content. PrincipleToxicity Principle = "toxicity" // PrincipleBrandSafety evaluates content alignment with brand guidelines. PrincipleBrandSafety Principle = "brand_safety" // PrincipleHonesty evaluates for deception or misleading content. PrincipleHonesty Principle = "honesty" // PrincipleRespect evaluates for respectful communication. PrincipleRespect Principle = "respect" // PrincipleSafety evaluates for dangerous instructions or content. PrincipleSafety Principle = "safety" // PrinciplePrivacy evaluates for privacy violations. PrinciplePrivacy Principle = "privacy" // PrincipleCustom is a user-defined principle. PrincipleCustom Principle = "custom" )
type Rule ¶
type Rule struct {
Principle Principle `json:"principle"`
Threshold float64 `json:"threshold,omitempty"` // score threshold (0.0-1.0)
Categories []string `json:"categories,omitempty"` // e.g., toxicity subcategories
Guidelines []string `json:"guidelines,omitempty"` // brand guidelines as text
Config map[string]any `json:"config,omitempty"`
}
Rule defines a specific value enforcement rule.
type Store ¶
type Store interface {
CreateValues(ctx context.Context, v *Values) error
GetValues(ctx context.Context, vID id.ValueID) (*Values, error)
GetValuesByName(ctx context.Context, appID, name string) (*Values, error)
UpdateValues(ctx context.Context, v *Values) error
DeleteValues(ctx context.Context, vID id.ValueID) error
ListValues(ctx context.Context, filter *ListFilter) ([]*Values, error)
}
Store defines persistence operations for values.
type Values ¶
type Values struct {
shield.Entity
ID id.ValueID `json:"id" bun:",pk"`
Name string `json:"name" bun:",notnull"`
Description string `json:"description,omitempty"`
AppID string `json:"app_id" bun:",notnull"`
TenantID string `json:"tenant_id,omitempty"`
Rules []Rule `json:"rules,omitempty" bun:"type:jsonb"`
Severity string `json:"severity,omitempty" bun:",default:'warning'"` // how seriously violations are treated
Action string `json:"action" bun:",notnull"` // "block", "flag", "warn"
Enabled bool `json:"enabled" bun:",notnull,default:true"`
Metadata map[string]any `json:"metadata,omitempty" bun:"type:jsonb"`
}
Values represents the moral compass and ethical guidelines. Values are the fourth layer of the safety engine. They evaluate content against ethical principles with configurable thresholds.
Source Files
Directories
¶
| Path | Synopsis |
|---|---|
|
Package principles holds the built-in detectors of the values layer: a curated lexicon for the safety principle (destructive infrastructure commands, the shell ones taken from instinct/detection) and the toxicity principle (abuse and threats aimed at people), and the score that turns pattern matches into a violation score.
|
Package principles holds the built-in detectors of the values layer: a curated lexicon for the safety principle (destructive infrastructure commands, the shell ones taken from instinct/detection) and the toxicity principle (abuse and threats aimed at people), and the score that turns pattern matches into a violation score. |