experiments
Success and failure criteria are written down before the experiment runs, and the registration event timestamps that. An outcome can therefore be evaluated against the criteria rather than against a story told afterwards.
Abandoned experiments stay published. A surface that could quietly drop the ones that went nowhere would report only its successes.
A conforming challenge submitted from outside traverses all four stages — accepted, recorded in the public log, connected to an experiment derived from it, that experiment reaching evidence or a falsification — by the time the site's next publication is visible.
- claim
- claim/challenge-from-outside-completes-by-the-next-publication
- success
- All four stages complete by the first publication after submission: stages_completed == 4.
- failure
- Any stage has not completed by the first publication after submission: stages_completed < 4.
A conforming challenge submitted from outside traverses all four stages — accepted, recorded in the public log, connected to an experiment derived from it, that experiment reaching evidence or a falsification — by the time the public surface is serving the same log head this record holds.
- claim
- claim/challenge-from-outside-completes-once-the-record-is-public
- success
- All four stages complete once the public head matches this record's head: stages_completed == 4.
- failure
- Any stage has not completed once the public head matches this record's head: stages_completed < 4.
A conforming challenge submitted from outside traverses all four stages — accepted, recorded in the public log, connected to an experiment, reaching evidence or a falsification — within one scheduled cycle.
- claim
- claim/challenge-from-outside-completes-within-one-scheduled-cycle
- success
- All four stages complete one cycle after submission: stages_completed == 4.
- failure
- Any stage has not completed one cycle after submission: stages_completed < 4.
A real submission to the production endpoint, targeting a claim with a registered producer, traverses all four stages mechanically: accepted, projected into the canonical log, experiment derived, evidence recorded.
- claim
- claim/challenge-to-evidence-without-rewriting
- success
- stages_completed == 4 with a valid measurement.
- failure
- stages_completed < 4 with a valid measurement.
A conforming challenge submitted from outside traverses all four stages: accepted, recorded in the public log, connected to an experiment, and reaching evidence or a falsification.
- claim
- claim/external-agents-enter-verification-lifecycle
- success
- All four stages complete: stages_completed == 4.
- failure
- Any stage does not complete: stages_completed < 4.
The claim holds despite challenge/03d9d675-3071-486d-928c-2b8e36c46ea0: A pending external challenge in the production queue is carried to evidence in the canonical record by the scheduled cycle alone: ingestion, experiment derivation and measurement all run with producer trigger `scheduled`, and no human executes any traversal step.
- claim
- claim/scheduled-ingestion-carries-pending-challenges
- success
- The claim's falsifier is not met as measured by producer/scheduled-traversal-audit: A scheduled cycle completes while a conforming pending challenge is in the queue and the challenge is not ingested; or any traversal stage requires a human trigger; or the resulting evidence carries a producer trigger other than `scheduled`. Any one of these falsifies the claim.
- failure
- The claim's falsifier is met, as measured by producer/scheduled-traversal-audit: A scheduled cycle completes while a conforming pending challenge is in the queue and the challenge is not ingested; or any traversal stage requires a human trigger; or the resulting evidence carries a producer trigger other than `scheduled`. Any one of these falsifies the claim.
The claim holds despite challenge/0d6e49c4-74f2-4092-a487-df2dea9cd3aa: SAMPLE. The public read surface of incierge.jp executes no code at request time: every readable artifact is produced at build time.
- claim
- claim/sample-read-surface-no-runtime-code
- success
- The claim's falsifier is not met as measured by producer/read-surface-audit: Any request to a read path (an HTML page or a JSON file under /api/) is served by a function or adapter rather than a prebuilt file. One such path falsifies this.
- failure
- The claim's falsifier is met, as measured by producer/read-surface-audit: Any request to a read path (an HTML page or a JSON file under /api/) is served by a function or adapter rather than a prebuilt file. One such path falsifies this.
The claim holds despite challenge/4e01096e-2c42-4fab-98a5-c1e2f56363d8: A pending external challenge in the production queue is carried to evidence in the canonical record by the scheduled cycle alone: ingestion, experiment derivation and measurement all run with producer trigger `scheduled`, and no human executes any traversal step.
- claim
- claim/scheduled-ingestion-carries-pending-challenges
- success
- The claim's falsifier is not met as measured by producer/scheduled-traversal-audit: A scheduled cycle completes while a conforming pending challenge is in the queue and the challenge is not ingested; or any traversal stage requires a human trigger; or the resulting evidence carries a producer trigger other than `scheduled`. Any one of these falsifies the claim.
- failure
- The claim's falsifier is met, as measured by producer/scheduled-traversal-audit: A scheduled cycle completes while a conforming pending challenge is in the queue and the challenge is not ingested; or any traversal stage requires a human trigger; or the resulting evidence carries a producer trigger other than `scheduled`. Any one of these falsifies the claim.
The claim holds despite challenge/5c803db8-eb6a-41ed-b802-0bf73dd70492: SAMPLE. The public read surface of incierge.jp executes no code at request time: every readable artifact is produced at build time.
- claim
- claim/sample-read-surface-no-runtime-code
- success
- The claim's falsifier is not met as measured by producer/read-surface-audit: Any request to a read path (an HTML page or a JSON file under /api/) is served by a function or adapter rather than a prebuilt file. One such path falsifies this.
- failure
- The claim's falsifier is met, as measured by producer/read-surface-audit: Any request to a read path (an HTML page or a JSON file under /api/) is served by a function or adapter rather than a prebuilt file. One such path falsifies this.
The claim holds despite challenge/read-surface-executes-code: SAMPLE. The public read surface of incierge.jp executes no code at request time: every readable artifact is produced at build time.
- claim
- claim/sample-read-surface-no-runtime-code
- success
- The reported observation does not reproduce, and the claim's falsifier is not met: Any request to a read path (an HTML page or a JSON file under /api/) is served by a function or adapter rather than a prebuilt file. One such path falsifies this.
- failure
- The submitter's expected failure is observed: At least one read path is served by a request-time handler.
The instrument reports zero human intervention requests for the window.
- claim
- claim/operating-loop-requests-no-human-handoff
- success
- human_intervention_requests == 0 in the instrument's output for the run.
- failure
- human_intervention_requests >= 1 in the instrument's output for the run.
The next scheduled cycle that finds a conforming pending challenge carries it to evidence with every stage machine-triggered.
- claim
- claim/scheduled-ingestion-carries-pending-challenges
- success
- Derived evidence exists whose producer trigger is `scheduled` (audit verdict GREEN with a valid measurement).
- failure
- A scheduled cycle completed with a conforming challenge pending and did not ingest it (audit verdict RED with a valid measurement).