Pattern-History Authorization — Detail

Load this after pattern_history_status.py returns and before consuming any catalog-derived historical field. Phase 0 runs the script; this file defines what its output authorizes.

The status payload

"{python_path}" "{speaker_toolkit_root}/skills/presentation-creator/scripts/pattern_history_status.py" \
  path/to/speaker-profile.json path/to/rhetoric-style-summary.md

Use - for the profile path in summary-only mode. The command emits {history_enabled, history_source, profile_schema_version, scored_talk_count, eligible_talk_count, opportunity_rows_available, classification_fields_available, available_classification_domains, policy_semantic_sha256, reason_codes, reasons, warning}.

history_enabled means at least one policy-bound domain is available. It is not permission to consume every derived field. In JSON, require membership in available_classification_domains for each use. Python consumers call domain_available(domain).

Domain contracts

[NEW] is exactly mastery_levels.never_tried / never_used_patterns. It is never first detection in the newest talk, and never a not_yet_observed classification.

A recurrence claim may be available without a trend claim.

Source selection

A non-null history_source selects the sole catalog-history input. Use the emitted value without reproducing source-selection logic and without merging inputs. Surface a disabled result's warning verbatim and recommend profile regeneration.

Continue using independent non-pattern fields regardless of history status: pacing, visual rules, presentation modes, infrastructure, publishing config, and confirmed intents.

Schema tiers

• Stored profile schemas v1/v2/v3 remain readable for non-pattern fields only.
• Schema v4 is occurrence-only.
• Schema v5 binds derived classifications to a versioned policy.

Exact occurrence rows may remain auditable when opportunity_rows_available is true, but that status never authorizes a classification.

Suppress each catalog-derived historical field when its required domain is absent. Do not collapse the available domains behind the global history flag.

Top-level recurring issues and badges in schema-v4/v5 profiles remain usable only when their entries explicitly declare source_lane: "non_pattern". Legacy or ambiguous entries do not authorize history. Current-taxonomy scans of the new outline remain enabled.

Summary-only mode

If no profile exists, run in summary-only mode: default guardrail thresholds (1.5 slides/min, 45% Act 1 cap) and ask for template/publishing data interactively.

Section 15 v3 classifications are usable only when its uniquely delimited current block passes:

"{python_path}" "{speaker_toolkit_root}/skills/vault-profile/scripts/section15_pattern_history.py"

plus the same strict pattern-profile assessment. Require the relevant domain just as for profile history. Section 15 v2 is occurrence-only. Ordinary, stale, or occurrence-only Section 15 data authorizes taxonomy-only recommendations, never speaker-history claims.

Comparing two profiles

Compare pattern catalog fingerprints and scoring schemas first. A mismatch is a generation reset — do not call cross-generation pattern differences improvements or regressions.

Raw scores are also incomparable when the baseline reports an unavailable or changed opportunity_coverage_identity.

Within one catalog/scoring generation, a changed policy_semantic_sha256 is a classification-comparison reset: do not describe changed tiers, recurrence classes, combinations, or trends as speaker improvement or regression across that boundary.