H3 Context Validator:
The H3 Context Validator audits prompt structure and hard constraints without changing or repairing the prompt text. It provides a checking stage in an H3 context workflow, helping you identify structural or constraint-related problems before continuing with later processing. Its approach is to examine a planned context alongside the corresponding prompt document through the canonical validation pipeline, preserving any failures rather than silently correcting them. This makes the audit useful when you need to confirm that a prompt follows the workflow’s requirements while retaining control over every revision. Validation concerns structure and constraints; it should not be treated as a guarantee of artistic quality or successful image generation.
H3 Context Validator Input Parameters:
plan
Required input of type H3_CONTEXT_PLAN. This is the typed plan under audit: a structured workflow object rather than a plain text instruction. The validator examines it together with the prompt document to audit the pair through the canonical context pipeline. Connect the plan associated with the document you intend to check so that the audit uses the appropriate planning context. An unrelated or incompatible object can prevent validation from completing or lead to diagnostics, depending on the pipeline’s checks. No default value, numeric minimum, numeric maximum, or selectable options are defined; this input is supplied through a compatible connection.
prompt_document
Required input of type H3_PROMPT_DOCUMENT. This is the rendered or explicitly parsed prompt document to be audited alongside the plan. It must be a structured prompt-document object, not simply text pasted into an unrelated text socket. The document supplies the prompt structure being checked, and the validator leaves its text unchanged even when it finds problems. Connect the document corresponding to your chosen plan; after editing a prompt, render or parse it again upstream as appropriate before rerunning the audit. No default value, numeric minimum, numeric maximum, or selectable options are defined.
H3 Context Validator Output Parameters:
validation
Output of type H3_VALIDATION_RESULT. This structured result records the outcome of the audit and provides diagnostics for understanding detected problems. The supplied validation lifecycle code constructs a result containing a validation identifier, schema version, target document identifier, status, diagnostics, and validator version. Its status uses the PASSED or FAILED validation states, depending on whether validation detects failure. Use the status to determine whether the document passed the checks, and inspect the diagnostics for details rather than treating the result as a simple text message. A failed validation does not mean the prompt has been repaired; detected problems remain visible for you to address.
validated_report
Output of type H3_CONTEXT_REPORT. This is the canonical context report carrying the terminal validation state, making the audited report available to compatible downstream nodes. Here, “validated” means that the report has been through validation, not that it necessarily passed: the report can preserve a failed state. This output is useful when later workflow stages need the canonical report together with its validation outcome rather than only the standalone validation result. It does not represent an automatically corrected prompt, and you should check its validation state before treating the audited content as ready for further use.
H3 Context Validator Usage Tips:
- Place this node after the workflow has produced both a plan and a rendered or explicitly parsed prompt document. Connect compatible typed outputs rather than raw text.
- Keep the plan and document synchronized. If you change the plan or revise the prompt upstream, regenerate the affected objects and run validation again.
- Inspect
validationfor the outcome and diagnostics; usevalidated_reportwhen a downstream node requires anH3_CONTEXT_REPORTconnection. - Treat a failed result as a request for review, not as an automatic repair step. Correct the reported issue upstream and repeat validation.
- There are no exposed strictness, repair, or performance settings on this node. Effective configuration consists of supplying the correct matching objects and responding to the resulting diagnostics.
- Connect the outputs to a compatible downstream consumer when you want the audit included in an executing workflow. The class declares
OUTPUT_NODE = False, so it is not designated as a terminal output node.
H3 Context Validator Common Errors and Solutions:
Required input is missing: plan or prompt_document
- Explanation: Both connections are required by the node interface. The preliminary input-checking method temporarily allows unresolved values, but this does not make either input optional during execution. The supplied source does not show the exact missing-input error text.
- Solution: Connect both required sockets to compatible upstream outputs and ensure those upstream nodes can execute successfully.
Incompatible object supplied to a typed input
- Explanation: The validator expects an
H3_CONTEXT_PLANand anH3_PROMPT_DOCUMENT. Plain text or another object type is not a substitute. The exact diagnostic for incompatible data is not included in the supplied source. - Solution: Use upstream nodes that produce the required socket types. If you start with prompt text, use the appropriate workflow step to render or explicitly parse it into a prompt document before validation.
Validation status: FAILED
- Explanation: The validation lifecycle supports a failed state when its checks detect a failure. This is an audit outcome, not necessarily an execution exception, and the node deliberately does not correct the underlying problem.
- Solution: Read the diagnostics in
validation, revise the relevant upstream content, and rerun the node. Do not assume that receiving avalidated_reportmeans the audit passed.
PipelineNodeError carrying validation diagnostics
- Explanation: The node catches
CanonicalContextPipelineErrorand raisesPipelineNodeErrorwith the original diagnostics. This indicates that the canonical pipeline reported an error; the visible message depends on those diagnostics, whose exact wording is not provided. - Solution: Follow the diagnostic details to identify the affected upstream object or requirement. Check that both inputs are valid, compatible workflow objects, correct the reported issue, and execute again.
Problems remain after validation
- Explanation: Unchanged prompt text is expected behavior, not a repair failure. The node explicitly audits without mutating or repairing the document.
- Solution: Make corrections in the upstream planning, rendering, parsing, or editing stage responsible for the content, then supply the updated objects to the validator.
