Every Insider document separates four things that are usually conflated: what was observed, whether the observation completed, how certain it is, and what it is allowed to mean. This page gives the vocabulary and the reading order so an agent does not promote one axis into another.
Reading order
- Schema and identity. Check
schemais the one you expected and, for content-addressed documents, thatidequals"sha256:" <> integrity.content_digestand that the appropriatevalidate_with_*function accepts the document with its exact inputs. Standalone validation proves structure and content identity only. - Operational status. Did the operation run to completion? Protocol
results:
ok,error,timeout,cancelled,unsupported. Confirmation records:operational_statusofcompleted,cancelled,error,timeout. CLI exits: 0, 1, 124, 130, 3. A non-completed operation can still carry a true partial finding, but never a success. - Domain outcome. The thing the operation was asked about. Oracles:
changed/unchanged/inconclusive(orincreased/unchanged/decreased/inconclusive). Confirmation:matching_failure,nonreproduced,repaired,unrelated_failure,infrastructure_failure. Verification:verifiedorinvalid. Minimizer and bisection:completed,exhausted,inconclusive, with areason. - Bounds and truncation.
truncated,truncation_reasons,dropped,requested_limitsversuseffective_limits,consumption,bounds,*_truncated,*_observed. A bound that was hit narrows every conclusion to what fit inside it. - Cleanup.
not_required,complete,partial,uncertain,unknown, or for searchescomplete/incomplete/unknown. Onlycompleteandnot_requiredmean nothing was left behind. - Confidence and uncertainty.
confidence(certain,high,medium,low,unknown) and theuncertaintylist of named facts the component could not establish. Read every entry; they are the component telling you where its claim stops. - Controls and limitations. Fixed per schema.
controlssay what the artifact cannot do;limitationssay what it cannot know. They do not vary per run and are not a verdict; they are the contract you are reading under.
Words that are not success
| Word | What it means | What it does not mean |
|---|---|---|
candidate | a static signal worth investigating | a defect, a cause, an impact |
syntactic | an identity read from source without macro expansion | a runtime fact |
changed | a global differed between the two captures of one run | which test changed it, that it matters, that it reaches production |
inconclusive | a bound or read failure prevented comparison | unchanged, clean, or safe |
operational_failure | the artifact never observed anything | any statement about the target |
nonreproduced | the scenario ran and the expected finding was not observed | the defect is absent; the scenario recipe ran, which is all completion means |
matching_failure | the expected finding was observed on a disposable target | causality, production impact, or a repair direction |
repaired | a configured positive oracle reported repair | repair of anything outside the declared bounded scenario |
failed (regression) | two fresh, distinct runs both reproduced on the same baseline | that the record's asserted execution occurred; a trusted runtime must have produced the live records |
completed (minimizer) | no single remaining whole step can be deleted | a globally smallest reproducer |
adjacent_boundary_isolated | one adjacent nonreproduced/reproducing pair on the search path | earliest reproducing target, or cause |
not_proven | the search exhausted its budget; the best bracket so far | failure of the search |
verified | both after-runs passed with exact provenance and complete cleanup | generalized bug freedom; proof of the declared scenario only |
invalid | at least one verification reason | that the candidate is wrong; it may be a cleanup or provenance problem |
declared_unverified | candidate bytes bound to a handoff | anything about the candidate's effect |
eligible (proposal) | the evidence graph recomputes | permission to publish; publication=blocked stands |
provider_reported_created | what the test provider said | independently verified remote state |
passed (load) | every structural gate held on this host | a production budget or capacity |
no_canary_hits | no declared token surfaced from the supplied items | application coverage or redaction approval |
mock: true, not_attempted | the observability mapping is deterministic data | delivery or integration |
Fingerprints and identities
sha256:identities are content, not authenticity. Two equal digests mean equal bytes; they do not say who produced them. Trust comes from the configured adapter, provider, or host that validated the inputs.- Oracle
hmac-sha256:fingerprints use a random per-run key that is never exported. Compare them only inside the report they appear in. - Provenance has five components (
code,configuration,dependencies,runtime,target), each available, unavailable with a reason, or redacted under a named policy fingerprint. Anunavailablecomponent is not a mismatch; it is a documented gap, and the manifest keeps unavailable and redacted distinct. contextin an oracle report is whatever the caller passed in. It is echoed, never verified, and every report says so.
Binding evidence to the world
No oracle, inventory, or load report binds itself to the test you ran, the
revision you ran it on, or the ticket you are working. The inventory gives
you source_fingerprint and candidate id; the oracle gives you context;
the load report binds git_sha and tree state but only for the host it ran
on. You must record the command, the exact selected test, the revision, and
the inventory identity next to the report. caller_must_bind_test_and_revision
appears in every oracle report for this reason.
When to stop
Stop and report rather than proceed when:
- a
70exit, a missing oracle line, orinfrastructure_failureappears: the environment is broken, not clean; - an inventory's
completenessis anything butcomplete(exits3to5) and you were about to treat a missing candidate as absence of risk; scanner_revisiondiffers between inventories you are comparing;cleanupis anything butcompleteornot_requiredon a step whose next step assumes a clean executor;uncertaintynames the exact thing your next inference depends on;- a
syntacticidentity has not yet been validated against the real module environment; - a record validates standalone but you have not run its
validate_with_*with the exact inputs.
An unconfirmed finding does not authorize a repair, and a verified repair does not authorize publication. Those are the two sentences the whole chain is built to keep true.