name: constraint-validator description: A skill to verify that the system is adhering to its own epistemic and operational rules.

Constraint Validator Skill

Status: Operational
Skill Category: Validation (Semantic)

1. Skill Purpose

The constraint-validator performs semantic logic verification on system artifacts (Narratives, Signals, Decisions). It ensures that the truth contained in these artifacts following the system's own logical constraints (e.g., No Future Leaks).

2. Invocation Contract

Standard Grammar

Invoke constraint-validator
Mode: <VERIFY>
Target: <path/to/artifact | artifact_json>
ExecutionScope:
  mode: <narrative | signal | decision>
Options:
  strict: <enabled | disabled>

3. Supported Modes & Selectors

• VERIFY: This is the only supported mode. It provides a binary PASS/FAIL judgment on the target object.
• Selectors (Scopes):
  • narrative: Checks for regime alignment, genesis fields, and temporal integrity.
  • signal: Verifies signal-to-regime compatibility.
  • decision: Traces decisions back to roots and verifies "Why"/ "Stop" fields.

4. Hook & Skill Chaining

• Chained From: Triggered by design-build-harness after artifact-producing tasks.
• Chained To: Fed into the audit-log-viewer.

5. Metadata & State

• Inputs: JSON/Markdown artifacts, active regime/factor state.
• Outputs: Binary status, violation report.

6. Invariants & Prohibitions

1. Read-Only: NEVER alters the target data.
2. No Partial Credit: Verification is strictly binary; a single violation fails the object.
3. Temporal Consistency: Future timestamps or "lookahead" data in past artifacts are terminal failures.

7. Example Invocation

Invoke constraint-validator
Mode: VERIFY
Target: data/narratives/2026-01-24__narrative.json
ExecutionScope:
  mode: narrative
Options:
  strict: enabled