Skip to content

Harden the owner contract, evals, and release pipeline (3.0.1) - #96

Merged
mzored merged 4 commits into
mainfrom
v3-hardening
Sep 3, 2026
Merged

Harden the owner contract, evals, and release pipeline (3.0.1)#96
mzored merged 4 commits into
mainfrom
v3-hardening

Conversation

@mzored

@mzored mzored commented Sep 3, 2026

Copy link
Copy Markdown
Owner

Release 3.0.1. Full notes in CHANGELOG.md.

What changed, by group

  • Owner contract (plugins/skiphow/skills/skiphow/): authority has a provenance (owner and trusted host policy versus repository instruction files); a commit is made only where its path is known to be safe; a read-only review never becomes repair; completion is relative to the authorized destination; earlier runs' branches are inspected, not cleaned; delegation states an authority boundary, defers to the kernel's isolation rule instead of restating it, and routes model and effort by consequence and complexity; writing-for-agents drops unsupported generalizations. Covered by tests/test_contract.py.
  • Behavioral corpus (evals/, docs/evidence.md): twenty-four cases over eighteen fixtures and five arms, each linked into the shipped contract, with a semantic validator that rejects a case no arm could satisfy. Two maintainer-only cases removed.
  • Release pipeline and host matrix (.github/workflows/, SECURITY.md, scripts/check_hosts.py): validate (contents: read) runs checks and the pinned Codex validator and refuses a tag not matching VERSION or not contained in origin/main; publish-release (contents: write) runs no repository code; deploy-site deploys the validated artifact. Manual Pages deploy requires an exact tag or SHA. SECURITY.md carries the per-capability host matrix. Covered by tests/test_workflows.py and tests/test_hosts.py.
  • Docs, validator, version: docs/decisions.md indexes live decisions; the deterministic validator stops failing on editorial shape and validates the hook on its safety properties; VERSION and both manifests at 3.0.1.

Checks run locally

  • python scripts/check.py: pass
  • python scripts/check.py --pytest -q tests: 236 passed
  • git diff --check: clean

Repository settings changed for the release

  • GitHub Pages build_type was already workflow.
  • The github-pages environment allowed only branch main; a v* tag deployment policy was added so the tag-triggered deploy-site job can run.

UNVERIFIED

  • No model run has been made on this or any 3.x package; every behavioral claim is the text's.
  • Clean Codex install and the Codex validator locally; the Codex validator runs in the validate job of the release workflow.

🤖 Generated with Claude Code

mzored and others added 4 commits September 4, 2026 01:49
Delegation defers to the kernel's isolation rule instead of restating it,
maintainer-only calibration advice leaves the runtime text, and the
contract tests assert the boundaries rather than the wording.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Every case states per-arm activation expectations and links into the
shipped contract; the evidence page records the corpus as UNVERIFIED.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…er capability

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
@mzored
mzored merged commit 179f344 into main Sep 3, 2026
1 check passed
@mzored
mzored deleted the v3-hardening branch September 3, 2026 21:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant