spanforge.sdk.policy
Runtime policy engine for SpanForge GA governance controls.
SFPolicyClient
from spanforge.sdk import sf_policy
sf_policy is the control-plane layer used by sf_explain, sf_scope, sf_rbac, sf_rag, and sf_lineage.
Core Lifecycle
load_bundle(bundle)
Load a RuntimePolicyBundle into the local registry.
activate(environment, policy_id, version, activated_at)
Make one loaded bundle active for an environment.
deactivate(environment, deactivated_at)
Remove the active bundle for an environment.
promote(...)
Clone a bundle version from one environment into another.
Runtime Evaluation
evaluate(...)
Evaluate one service/control pair against the active environment bundle and emit a signed decision.
Supported decision actions:
allowallow+logredactblockhuman_review
Replay and Simulation
simulate(...)
Run a candidate bundle against one event without changing production state.
replay(...)
Replay historical events through a candidate bundle and return a summary with simulation records.
compare_policies(...)
Summarize action changes between baseline and candidate bundles.
Review Loop
record_review(...)
Record analyst feedback on a decision with one of:
false_positivetrue_positiveneeds_tuning
suggest_threshold(...)
Derive a threshold hint from reviewed decisions.
Signed Audit Records
sf_policy writes the following schema keys into sf_audit:
spanforge.policy.lifecycle.v1spanforge.policy.decision.v1spanforge.policy.simulation.v1spanforge.policy.replay.v1spanforge.policy.comparison.v1spanforge.policy.review.v1
Ready to instrument your AI pipeline?