Source: openagent-sdk/crates/openagent-skills-policy/rust/src/engine.rs. SHA-256: 9e55ac1b6524d85519315932929f89032d22db8f361634fcd7adc6b13a394abe.

This source reference follows declared modules and preserves feature attributes. It includes public declarations and implementation methods in those modules. Private-module exports and trait resolution still require the compiler; not every declaration is a crate-root import. Function bodies and constant values are omitted. Source comments describe their implementation context and are not a production deployment claim.

engine::InvocationContext

Context passed into can_invoke / record_invocation describing who is attempting to use a skill, when, and with what arguments.

#[derive(Debug, Clone)]
pub struct InvocationContext {
/// DID of the invoking agent.

pub agent_did: Did,
/// Session ID (opaque string — matches `arsenal-core::SessionId` shape).

pub session_id: String,
/// Arguments that will be passed to the skill.

pub arguments: serde_json::Value,
/// Wall-clock time of the attempted invocation (UTC).

pub invoked_at: OffsetDateTime,
/// Whether the caller has already collected human-in-the-loop consent.

pub consent_granted: bool
}

Source line: 31.

engine::InvocationContext::new

Convenience constructor. consent_granted defaults to false.

pub fn new(
        agent_did: Did,
        session_id: impl Into<String>,
        arguments: serde_json::Value,
    ) -> Self;

Source line: 46.

engine::InvocationContext::with_consent

Mark consent as granted (e.g. after a HITL approval flow).

#[must_use]
pub fn with_consent(mut self) -> Self;

Source line: 62.

engine::InvocationContext::with_invoked_at

Override the invocation timestamp (primarily for testing).

#[must_use]
pub fn with_invoked_at(mut self, t: OffsetDateTime) -> Self;

Source line: 69.

engine::SkillsPolicy

Governance engine for agent skill invocations.

pub struct SkillsPolicy {

}

Source line: 76.

engine::SkillsPolicy::from_yaml

Parse a policy YAML document.

Errors

Returns [SkillsPolicyError::InvalidPolicy] if the YAML is malformed or fails structural validation.

pub fn from_yaml(yaml: &str) -> Result<Self>;

Source line: 101.

engine::SkillsPolicy::from_file

Load a policy YAML document from a file path.

Errors

Returns [SkillsPolicyError::Io] on read failure, or any error from [SkillsPolicy::from_yaml] on parse/validation failure.

pub fn from_file(path: &Path) -> Result<Self>;

Source line: 119.

engine::SkillsPolicy::agent

The DID of the agent this policy applies to.

#[must_use]
pub fn agent(&self) -> &Did;

Source line: 129.

engine::SkillsPolicy::has_rule

Whether the policy has a per-skill rule for skill.

#[must_use]
pub fn has_rule(&self, skill: &str) -> bool;

Source line: 135.

engine::SkillsPolicy::can_invoke

Read-only check: can ctx invoke skill right now?

This runs every policy dimension except the rate-limit advance. A successful return does not consume any rate-limit tokens — use record_invocation for that.

Errors

Returns a specific [SkillsPolicyError] variant per failed dimension.

pub fn can_invoke(&self, skill: &str, ctx: &InvocationContext) -> Result<()>;

Source line: 147.

engine::SkillsPolicy::record_invocation

Authorize and record an invocation.

This runs can_invoke, then checks and advances the rate limiter, then appends a [Receipt] to the hash-chained audit log. On any failure no state changes.

Errors

Returns the first dimension that failed. On rate-limit failure the caller gets used and max values to surface a friendly error.

pub fn record_invocation(
        &mut self,
        skill: &str,
        ctx: &InvocationContext,
    ) -> Result<Receipt>;

Source line: 213.

engine::SkillsPolicy::audit_chain

Borrow the audit chain for inspection (verification, export).

#[must_use]
pub fn audit_chain(&self) -> &AuditChain;

Source line: 281.

engine::SkillsPolicy::reset_rate_limit

Reset the rate-limit state for a single skill.

pub fn reset_rate_limit(&mut self, skill: &str);

Source line: 286.

engine::arsenal_error_from

Used by integration with Arsenal: convert a SkillsPolicyError into an arsenal_core::ArsenalError. Kept behind the engine module so that downstream crates can depend on this without pulling the whole module graph.

pub fn arsenal_error_from(err: &SkillsPolicyError) -> arsenal_core::ArsenalError;

Source line: 370.