[borek504]

Purpose

Make the existing component on main easy to try without copying a long README code block. This is one onboarding change, not a new validation protocol.

Four-file scope

• demo_execution_contract.py: runnable, standard-library-only synthetic demo showing a consistent report, a wrong attempt ID and a missing required test. It prints actual validator results, plain-language explanations and error codes.
• test_demo_execution_contract.py: seven focused regressions, including exact outcomes, isolated inputs, README result summaries, CLI execution and nonzero failure behavior even under python -O.
• README.md: replace the long inline example with a Quick Start and link to editable example inputs; keep the profile constraints and trust-boundary warnings.
• .github/workflows/tests.yml: retain the original focused test step and add demo regression tests plus the exact demo command.

The validator and original 33 contract tests are byte-for-byte unchanged. License and other existing files are unchanged. No new dependencies, package release, schema changes, qualification authority or merging of the separate candidates in #4/#10.

Try this PR before merge

git clone --branch docs-quickstart-demo --single-branch https://github.com/borek504/execution-evidence-contracts.git
cd execution-evidence-contracts
python -B demo_execution_contract.py
python -B -m unittest -v test_r14_execution_contract test_demo_execution_contract

The README's default-branch clone command is intended for after this PR is merged; for this unmerged change use the branch-specific command above.

Expected summaries:

[1] Consistent report: ACCEPTED (data only)
[2] Wrong attempt: REJECTED
[3] Missing required test: REJECTED

Every case prints qualification_authorized=False. Exit 0 means all three expected outcomes matched, including both rejections. Unexpected acceptance, the wrong reason or an authorizing result fails the demo.

Verification

Local Linux / Python 3.13.5: 33 existing + 7 demo tests PASS. Repeated from a fresh disposable directory containing only the needed public files; demo also passes under optimization. File inventory and source bytes remained unchanged during execution. All four new/changed file blobs match the locally tested copies.

Remote CI is pending verification at PR creation; local results are not a remote CI claim. These are synthetic examples: no workload described by the records, real observer, agent, external API or evidence-file write is executed. Only the synthetic report digest is refreshed in negative examples; expectations and observation bindings remain fixed. This is not a production evidence collector. Prepared with AI assistance; it does not satisfy the external review gate for #4.

[borek504]

CI verified for the submitted demo change

Head: 4a61e1d1b96a75ae4ca9424b59f795e352ec0851.

PR-triggered workflow 34965753418 has successful jobs on Python 3.12 and 3.13. Both job logs were read: each records 33 original contract tests + 7 demo tests PASS, followed by successful execution of python -B demo_execution_contract.py. Both demos report the expected accepted/rejected/rejected results and Demo checks: 3/3 matched the expected outcomes. No tests were skipped.

The hosted job logs identify clean checkout 4b5595e3533ce6ef9956eab5cf3f9c3517008432, GitHub's temporary PR merge ref, not a merge into main. Local tests also passed from a fresh directory containing only the necessary public files; source bytes and file inventory remained unchanged.

The four submitted blobs match locally tested bytes. Original validator blob 1dfd7a8eecefb1a67da2439d758728e24ea2dc4d and original test blob ab27ef95d92117b8446602e481c82beb7d95bdce are unchanged. This PR remains open for normal review; no merge or release has been performed. To try it now, use the branch-specific clone command in the PR description. The external review requirements for the separate candidates #4 and #10 are unaffected.

[borek504]

Final scoped maintainer-directed review

The maintainer explicitly approved closing this onboarding change and preparing, but not publishing, an experimental release draft.

Reviewed the four-file diff, full demo-test and README contents, and workflow. The example calls the actual validator, keeps expectations fixed in its negative cases, prints non-authorizing results, and uses explicit expected-outcome/error checks rather than optimization-removable assertions. Tests cover exact acceptance/rejection, unexpected acceptance, wrong rejection reason, authorizing output, independent fixture instances and the documented command. The two added CI steps preserve the original test command. The change adds no dependency or validation-protocol change.

The current compare shows exactly four files and one commit. Validator blob 1dfd7a8eecefb1a67da2439d758728e24ea2dc4d and original test blob ab27ef95d92117b8446602e481c82beb7d95bdce are unchanged. Re-read exact-head PR workflow 34965753418: completed success, both Python 3.12/3.13 jobs and all original-test/demo-test/demo steps successful. This is existing pre-merge CI, not a newly run local or post-merge suite.

No blocker found within this documentation/demo scope. Proceed with the authorized normal merge guarded by this head SHA, then verify main identity and its new push CI. This review is AI-assisted/internal, NOT an independent security review, and does not satisfy or waive the separate #4/#10 review gates. No release publication is authorized by this comment.

[borek504]

Completed: merged demo and verified main CI

PR #12 was merged, with explicit maintainer approval, as f0b98b7e8df01944ca2ccd6a38015865ff18b021. Its parents are the previous main a5c8e7b991e5bb975d9e3d2b6b4a45d8337a41ac and reviewed head 4a61e1d1b96a75ae4ca9424b59f795e352ec0851; merged tree 461ddd0cd2830972d463db1bb2d87d6e5386a8b7 matches the submitted candidate. The validator and original contract tests are unchanged.

New post-merge main CI, not merely the earlier PR run: 34968547115 is completed/success. Both actual job logs were read and identify the real main merge checkout. On hosted Ubuntu 24.04.5 / CPython 3.12.14 and 3.13.15, each job reports 33 contract tests + 7 demo tests PASS, no reported skips, then successful execution of python -B demo_execution_contract.py with all three expected outcomes. This is hosted-runner verification, not a new local clone result or independent security audit.

Proposed first release: text draft only — NOT published

Prepared tag name: v0.1.0-alpha.1.
Exact source target: f0b98b7e8df01944ca2ccd6a38015865ff18b021.
Title: v0.1.0-alpha.1 — Experimental execution-evidence validator.
Intended classification: pre-release.

This comment preserves the proposed release description; no native GitHub Release draft, release publication or tag was created. The available connector has release-read operations but no release-create action. Publication remains a separate maintainer decision; recheck the pinned target and open findings before it.

Prepared release description

Included

An intentionally small, Python-standard-library reference implementation for consistency checking across an execution record, expectations fixed before execution and separately supplied verifier observations. Includes r14_execution_contract.py, its 33 synthetic contract tests, the runnable three-case demo, seven demo tests, Quick Start, profile/security/contribution documentation and Apache-2.0 license.

PR #12 adds onboarding, not a new protocol: the validator and original contract tests are unchanged. There are no third-party Python dependencies or API-key requirements.

Try the exact source

Use Python 3.12 or 3.13 in a new checkout:

git clone https://github.com/borek504/execution-evidence-contracts.git
cd execution-evidence-contracts
git checkout --detach f0b98b7e8df01944ca2ccd6a38015865ff18b021
python -B demo_execution_contract.py
python -B -m unittest -v test_r14_execution_contract test_demo_execution_contract

These instructions use the existing commit, not the proposed unpublished tag. Use python3 where that is the name of the selected interpreter.

Expected summaries:

[1] Consistent report: ACCEPTED (data only)
[2] Wrong attempt: REJECTED
[3] Missing required test: REJECTED

All cases report qualification_authorized=False. Exit 0 and Demo checks: 3/3 matched the expected outcomes. mean the expected acceptance and BOTH rejections matched, not that all reports were accepted.

Verification

The post-merge main workflow linked above passed on hosted Linux with CPython 3.12.14 and 3.13.15: 40 test methods per interpreter plus the demo command. This does not qualify macOS/Windows or a production runtime. The seven demo tests include failure-gate verification under Python optimization.

Experimental boundaries

The historical jarvis-r14-* schema identifiers and current profile mappings are retained. The profile's python_version == "3.13.15" is a requirement on supplied evidence, not a measurement of the interpreter running the demo. See PROFILE.md.

Valid data are not proof of actual execution, authentic provenance or independent observation, isolation, termination, replay resistance, or permission to execute/qualify/release anything. A real integration must establish trusted input origins. The synthetic fixture is not an evidence collector. The alpha label gives no stable-API or production-readiness promise.

Result Envelope v2 (#4) and Atomic Evidence Store (#10) are NOT included. Their separate review gates are unaffected. Follow SECURITY.md and CONTRIBUTING.md when reporting a scoped finding, using an exact commit and synthetic reproducer, without private data or credentials.

Licensed under Apache-2.0. Prepared with AI assistance; internal review and CI are not independent third-party security acceptance.

[borek504]

First experimental source release published — 2026-09-15

Following explicit maintainer approval, v0.1.0-alpha.1 is now a real published GitHub pre-release, not the earlier text draft.

Read-back verification:

• Release ID 389131453; draft=false, prerelease=true; published at 2026-09-15T12:37:05Z.
• The actual tag ref points directly to approved source commit f0b98b7e8df01944ca2ccd6a38015865ff18b021, tree 461ddd0cd2830972d463db1bb2d87d6e5386a8b7.
• The release contains the existing validator, synthetic demo, tests and documentation. It excludes Result Envelope v2 (Candidate: characterize v1 and externally review Result Envelope v2 #4) and Atomic Evidence Store (Candidate: POSIX atomic evidence snapshot store #10); their review requirements are unaffected. No additional binary assets or package-registry publication.
• The publication workflow checked the exact successful main push CI 34968547115 and its required Python 3.12/3.13 test/demo steps before creating the release. This is not a new independent security review or new host qualification.

The connector did not expose a direct release-create action. Publication used a single-purpose GitHub Actions workflow on publish-v0.1.0-alpha.1-once, with repository Contents write and Actions read permissions for that job only, no checkout and no third-party action. Job 104383317691 in run 34969890496 completed successfully; the release and tag were then read back through the connector. The workflow file matched its reviewed local blob 18cb2e5b54ac4ec946cc859f8ad7f4a776249b44.

After verification, the one-shot workflow file was removed from the helper branch in commit 50a2a4f661d62e7bad33e0d2c5204f770f1a8e5f; its current tree is identical to the approved source tree. The workflow is not in main or the release. Main was read back unchanged at f0b98b7e8df01944ca2ccd6a38015865ff18b021. No repository-wide permission/protection settings were changed.

Alpha means experimental source for evaluation: no production-readiness, authentic-evidence, execution/isolation, replay-prevention or authorization guarantee. This supersedes only the earlier unpublished-release status, not the documented limitations.