Onboarding friction
Install, archive, or path confusion
Use this when tobi.exe is missing, shipped examples are not where expected, or checksum verification is unclear.
Docs / Support
Report Stage 1 issues in a reproducible way
Use this page to classify what happened, prepare a clean reproduction, and route support questions or workflow friction without widening the product claim.
The best support reports are narrow, exact, and explicit about whether the issue
happened in canon, golden, onboarding, or CI.
Issue classes
Classify the problem as narrowly as possible
Docs mismatch
Published guidance does not match the released line
Use this when the public docs omit a practical step or describe the Stage 1 surface inaccurately.
Diagnostics clarity
The validator rejects input, but interpretation is unclear
Use this when the code, message, or span is present but hard to read or map to the current docs.
Unexpected reject
Input appears to fail contrary to the released surface
Use this when the shipped sample or previously stable accepted input rejects unexpectedly.
Fixture mismatch
golden result diverges from expectation
Use this when canonical output, hash, code, or span mismatches the fixture expectation.
Workflow friction
Local success does not translate cleanly into CI
Use this when checkout, download, checksum, path, or runner setup blocks GitHub workflow evaluation.
Reproduction checklist
What to include in every useful report
- exact command run
- exact input file or fixture path
- whether shipped sample material or partner-owned material was used
- exact output received
- whether the issue happened in
canonorgolden - diagnostic code and span when present
- expected behavior versus actual behavior
- release tag, branch, or host details when relevant
Useful feedback
What high-quality Stage 1 feedback looks like
- reproducible rather than paraphrased
- narrow in scope rather than platform-wide
- explicit about whether the issue is onboarding, diagnostics, mismatch, or workflow friction
- grounded in exact output rather than interpretation alone
- one issue class per report whenever possible
Accepted input
If canon produced the expected output shape, do not file that as a validator failure.
Deterministic reject
If the validator rejected input with a stable code and span, classify it as expected reject, unexpected reject, or diagnostics clarity.
Scope misunderstanding
If the request is really about runtime, backend, or API expectations, treat it as product-scope misunderstanding rather than a Stage 1 defect.
Where to go next
Use the right page before escalating further
Diagnostics reference
Check deterministic codes, parser versus semantic rejection, and mismatch classes.
Open
GitHub Actions starter
Review the narrow workflow pattern before filing CI-specific friction.
Open
Guided evaluation details coming soon
Use the placeholder page for deferred evaluation details when the question is no longer a bug or support issue.
Open
Customer routing
Need help, workflow fit, or access updates?
Public docs should get most operators to a clean first evaluation. When the next step is a support conversation, CI integration question, or broader product inquiry, route it directly to Organetic.