Struct PromptMetadata 

pub struct PromptMetadata {

    pub model: String,
    pub provider: String,
    pub search_mode: String,
    pub renderer_mode: String,
    pub date: String,
    pub cwd: String,
    pub prompt_fragments: Vec<String>,
    pub docs_map: Vec<String>,
    pub context_sources: Vec<ContextSourceMeta>,
    pub tool_catalog_size: usize,
    pub tool_names: Vec<String>,
    pub skill_catalog_size: usize,
    pub skill_names: Vec<String>,
    pub diagnostics: Vec<String>,
    pub history_reuse: bool,
    pub prev_context_hash: Option<u64>,
    pub transcript_tail_size: usize,
    pub has_user_turn: bool,
}

Metadata for a single prompt turn, suitable for append-only JSONL storage.

This is the audit record: enough to reconstruct what was sent without storing the content itself. Full raw provider payloads are deliberately excluded because they can contain prompt text, repo content, and secrets.

Fields

model: String

Selected model name.

provider: String

Selected provider label.

search_mode: String

Application-owned web-search backend label.

renderer_mode: String

Renderer mode label.

date: String

Rounded date (YYYY-MM-DD) used for cache stability.

cwd: String

Workspace root path.

prompt_fragments: Vec<String>

Ordered prompt fragment names included in the system prompt.

docs_map: Vec<String>

Model-visible documentation entry point paths.

context_sources: Vec<ContextSourceMeta>

Metadata for each loaded AGENTS.md source (no content).

tool_catalog_size: usize

Number of tools in the catalog sent this turn.

tool_names: Vec<String>

Tool names in the catalog sent this turn.

skill_catalog_size: usize

Number of available Agent Skills exposed as metadata this turn.

skill_names: Vec<String>

Available skill names exposed as metadata this turn.

diagnostics: Vec<String>

Self-knowledge diagnostics visible for this turn.

history_reuse: bool

Whether history reuse was active for this turn.

prev_context_hash: Option<u64>

Content hash of the root AGENTS.md from the previous turn, if any.

transcript_tail_size: usize

Number of transcript entries included in the projected tail.

has_user_turn: bool

Whether the user turn was non-empty.

Implementations

impl PromptMetadata

pub fn from_bundle(bundle: &PromptBundle) -> Self

Extract prompt metadata from a PromptBundle for session storage.

This captures the structural metadata of the turn — model, search backend, context sources (hashes and truncation, not content), tool count, and transcript tail size. It does not store prompt text, AGENTS.md content, or provider request/response bodies.

Trait Implementations

impl Clone for PromptMetadata

fn clone(&self) -> PromptMetadata

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for PromptMetadata

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for PromptMetadata

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for PromptMetadata

impl PartialEq for PromptMetadata

fn eq(&self, other: &PromptMetadata) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for PromptMetadata

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for PromptMetadata