Skip to content

Latest commit

 

History

History
60 lines (49 loc) · 3.55 KB

AGENTS.md

File metadata and controls

60 lines (49 loc) · 3.55 KB

Repository Automation

Parent scope: repository.

Path Contract
workflows CI quality gates and tag-driven release ordering
ISSUE_TEMPLATE Numbered issue forms and chooser configuration
pull_request_template.md Default PR template and machine-parsed snapshot decision labels
scripts/check_snapshot_provenance.py Base/head Git-blob comparison and PR declaration validation
scripts/clean_wheel_agent_acceptance.py Supplied wheel acceptance in temporary home/CWD/venv
scripts/run_differential_acceptance.py Source-query 200-seed/v1-v2 acceptance with exact executed identities and coverage evidence
scripts/run_dataset_acceptance.py Required installed-wheel dataset test with fail-closed JUnit and artifact evidence

Provenance Guard

The guard compares the supplied base/head commits, not just working-tree content. Changes anywhere under src/pt_snap_cli/snapshot/, including AGENTS.md, trigger the PR decision requirement. Supply exactly one Snapshot provenance decision: with updated or no-update. updated requires new nonblank provenance content; no-update requires one specific Snapshot provenance no-update reason: and no provenance-file change. Placeholder text and duplicate declarations fail.

Existing provenance lines cannot be rewritten, reordered or deleted. This content check also runs on main pushes and tag releases. Preserve PR edited events so correcting a declaration can rerun CI.

Installed-Wheel Acceptance

The acceptance script audits a supplied wheel against source and validates the reviewed fixture manifest before installing or importing. It installs offline into a temporary venv with system dependencies available by default. Supplying --wheelhouse instead installs dependencies offline into a venv without system packages and verifies their installed paths too. It verifies that pt_snap_cli resolves inside the venv, clears source/catalog/focus/resource-limit overrides, disables user site packages, and uses temporary HOME/CWD for skill installation and import-created focus.

It then checks help, bundled catalog/install status, capabilities, trusted fixture import/reuse, overview and a peak query. This is executable CLI acceptance, not a live model evaluation. --events-per-slice 4096 adds native-v2 and compatibility-v1 sharded imports/reuse, directory/manifest discovery, installed helper/peak skill guidance, and real peak/cross-shard event parity with the standalone CLI. It requires at least two physical slices on the selected device and verifies that explicit dataset operations do not change focus. The script accepts only the expected committed snapshot_with_empty_cache.pkl path, rejects an unhydrated pointer, and runs the provenance gate itself; also run fixture provenance tests before acceptance. Its output explicitly distinguishes offline CLI execution, stored-baseline grading (not run here), and live-agent execution (not run).

Harness changes use pytest tests/test_clean_wheel_agent_acceptance.py plus an actual audited-wheel invocation; mocked guard tests alone are not acceptance.

Change Together

Keep guard labels, the default PR template, and workflow inputs aligned. Run pytest tests/test_governance.py tests/test_release_workflow.py from the repository root for governance/wiring changes. Read scripts before invoking them: the guard is read-only, while wheel acceptance creates an environment and loads trusted pickle data. Workflow/package changes also use the source packaging contracts.